Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Hiromichi FujisawaSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works530 from public data
- 2,390
- Dense Trajectories and Motion Boundary Descriptors for Action Recognition1,800
The MBH descriptor shows to consistently outperform other state-of-the-art descriptors, in particular on real-world videos that contain a significant amount of camera motion.
- 622
- Handwritten digit recognition: benchmarking of state-of-the-art techniques586
The results of handwritten digit recognition on well-known image databases using state-of-the-art feature extraction and classification techniques are competitive to the best ones previously reported on the same databases.
- 550
- 538
- 535
- Robust Classification with Convolutional Prototype Learning479
This paper argues that the lack of robustness for CNN is caused by the softmax layer, which is a totally discriminative model and based on the assumption of closed world (i.e., with a fixed number of categories), and proposes a novel learning framework called convolutional prototype learning (CPL).
- 455
- Deep Direct Regression for Multi-oriented Scene Text Detection432
A deep direct regression based method for multi-oriented scene text detection that achieves the F-measure of 81%, which is a new state-of-the-art and significantly outperforms previous approaches.
- Drawing and Recognizing Chinese Characters with Recurrent Neural Network363
Under the RNN framework, a conditional generative model with character embedding is proposed for automatically drawing recognizable Chinese characters and can be recognized by the discriminative RNN model with high accuracy.
- 353
- Character Recognition Systems: A Guide for Students and Practitioners351
This chapter discusses the development of Character Recognition, Evolution and Development, and some of the techniques used to achieve this goal, including Bayes Decision Theory, as well as some new methods based onributed graph matching.
- 339
- Handwritten digit recognition: investigation of normalization and feature extraction techniques332
A comparison of normalization functions shows that moment-based functions outperform the dimension-based ones and the aspect ratio mapping is influential and the comparison of feature vectors shows that the improved feature extraction strategies outperform their baseline counterparts.
- 329
- From System 1 to System 2: A Survey of Reasoning Large Language Models314
This survey provides an overview of the progress in foundational LLMs and the early development of System 2 technologies, exploring how their combination has paved the way for reasoning LLMs and explores promising directions for advancing reasoning LLMs.
- Semantic expansion using word embedding clustering and convolutional neural network for improving short text classification304
A unified framework to expand short texts based on word embedding clustering and convolutional neural network and semantic cliques via fast clustering is proposed, which validates the effectiveness of the proposed method on two open benchmarks.
- Online and offline handwritten Chinese character recognition: A comprehensive study and new benchmark302
Although directMap+convNet can achieve the best results and surpass human-level performance, it is shown that writer adaptation in this case is still effective and a new adaptation layer is proposed to reduce the mismatch between training and test data on a particular source layer.
- ICDAR2019 Robust Reading Challenge on Multi-lingual Scene Text Detection and Recognition — RRC-MLT-2019301
The dataset, the tasks and the findings of the presented RRC-MLT-2019 challenge are presented, which has 4 tasks covering various aspects of multi-lingual scene text.
- 238
- 226
- Data-Distortion Guided Self-Distillation for Deep Neural Networks212
A more elegant self-distillation mechanism to transfer knowledge between different distorted versions of same training data without the reliance on accompanying models, which can effectively improve the generalization performance of various network architectures.
- Arbitrary Shape Scene Text Detection With Adaptive Text Region Representation209
Recurrent neural network based adaptive text region representation is proposed for text region refinement, where a pair of boundary points are predicted each time step until no new points are found, and text regions of arbitrary shapes are detected and represented with adaptive number of boundary Points.
- 199
- Improving handwritten Chinese text recognition using neural network language models and convolutional neural network shape models193
Evaluating comprehensively neural network language models (NNLMs) and hybrid NNLMs in handwritten Chinese text recognition and replacing the baseline character classifier, over-segmentation, and geometric context models with convolutional neural network (CNN) based models.
- Semantic Clustering and Convolutional Neural Network for Short Text Categorization185
A novel method to model short texts based on semantic clustering and convolutional neural network is proposed, which first discovers semantic cliques in embedding spaces by a fast clustering algorithm, then multi-scale semantic units are detected under the supervision of semantic clique, which introduce useful external knowledge for short texts.
- 177
- Hybrid CNN and Dictionary-Based Models for Scene Recognition and Domain Adaptation168
By integrating the complementary information of MLR, CFV, and the CNN features of the fully connected layer, the state-of-the-art performance can be achieved on scene recognition and DA problems.
- Evaluation of prototype learning algorithms for nearest-neighbor classifier in application to handwritten character recognition162
Reviewing some prototype learning algorithms for NN classifier design and evaluating their performance in application to handwritten character recognition shows that the algorithms based on parameter optimization generally outperform the LVQ.
- Handwritten Chinese text line segmentation by clustering with distance metric learning159
A novel text line segmentation algorithm based on minimal spanning tree (MST) clustering with distance metric learning that is made robust to handle various documents with multi-skewed and curved text lines.
- Towards Robust Pattern Recognition: A Review157
A comprehensive review of research toward robust pattern recognition from the perspective of breaking three basic and implicit assumptions: closed-world assumption, independent and identically distributed assumption, and clean and big data assumption, which form the foundation of most pattern recognition models.
- Total-Text: toward orientation robustness in scene text detection154
A new scene text detection model is proposed as the baseline for Total-Text, namely Polygon-Faster-RCNN, and its ability to detect text of all orientations is demonstrated.
- 152
- 147
- A new benchmark on the recognition of handwritten Bangla and Farsi numeral characters141
Some results of handwritten Bangla and Farsi numeral recognition on binary and gray-scale images are presented and some implementation choices of gradient direction feature extraction, some advanced normalization and classification methods are compared.
- BlockQNN: Efficient Block-Wise Neural Network Architecture Generation140
This paper provides a block-wise network generation pipeline called BlockQNN which automatically builds high-performance networks using the Q-Learning paradigm with epsilon-greedy exploration strategy and proposes a distributed asynchronous framework and an early stop strategy.
- Model-based stroke extraction and matching for handwritten Chinese character recognition131
This method is able to obtain reliable stroke correspondence and enable structural interpretation and some structural post-processing operations are applied to improve the stroke correspondence.
- Practical Network Blocks Design with Q-Learning.127
This work provides a solution to automatically and efficiently design high performance network architectures by focusing on constructing network blocks, which can be stacked to generate the whole network.
- Classifier combination based on confidence transformation127
The experimental results justify that confidence transformation benefits the combination performance of either fixed rules or trained rules, and justify that the cascaded strategy is a right way of multiple classifier combination.
- Bag of Tricks for 3D MRI Brain Tumor Segmentation122
This paper reviews the different kinds of tricks applied to 3D brain tumor segmentation with DNN, and divides such tricks into three main categories: data processing methods including data sampling, random patch-size training, and semi-supervised learning, model devising methods including architecture devising and result fusing, and optimizing processes including warming-up learning and multi-task learning.
- 119
- 117
- Performance evaluation of pattern classifiers for handwritten character recognition112
The results indicate that pattern classifiers have complementary advantages and they should be appropriately combined to achieve higher performance.
- Handwritten Mathematical Expression Recognition via Paired Adversarial Learning109
This work proposes a novel paired adversarial learning method to learn semantic-invariant features of handwritten mathematical expressions, named PAL-v2, and adopts a convolutional decoder to alleviate the vanishing and exploding gradient problems of RNN-based decoder, and improves the coverage of decoding with a novel attention method.
- 96
- Fast kNN Graph Construction with Locality Sensitive Hashing96
This paper proposes an efficient algorithm for approximating kNN graphs, which has the time complexity of O(l(d+logn)n) only (d is the dimensionality and l is usually a small number) and is much faster than most existing fast methods.
- 96
- Online Japanese Character Recognition Using Trajectory-Based Normalization and Direction Feature Extraction95
An online Japanese character recognition system using advanced techniques of pattern normalization and direction feature extraction is described, and one-dimensional and pseudo two-dimensional normalization methods are compared.
- Pseudo two-dimensional shape normalization methods for handwritten Chinese character recognition95
A new pseudo 2D normalization method using line density projection interpolation (LDPI), which partitions the line density map into soft strips and generate 2D coordinate mapping function by interpolating the 1D coordinate functions that are obtained by equalizing the linedensity projections of these strips.
- 86
- 81
- 80
- 80
- 79
- A robust model for on-line handwritten japanese text recognition79
A robust context integration model for on-line handwritten Japanese text recognition based on string class probability approximation that can flexibly combine the scores of various contexts and is insensitive to the variability in path length is described.
- Common Sense Knowledge for Handwritten Chinese Text Recognition76
Three methods are introduced to supplement the standard n-gram language model and the results show that the adoption of common sense knowledge yields improvements in recognition performance, despite the reduced concept list hereby employed.
- Regularized margin-based conditional log-likelihood loss for prototype learning73
A new prototype learning algorithm based on the conditional log-likelihood loss (CLL), which isbased on the discriminative model called log- likelihood of margin (LOGM), which yields higher classification accuracies than the MCE, generalized learning vector quantization (GLVQ), soft nearest prototype classifier (SNPC) and the robust soft learning vectorquantization (RSLVQ).
- 72
- LG-CNN: From local parts to global discrimination for fine-grained recognition70
This paper proposes a jointly trained Convolutional Neural Network (CNN) architecture to solve the fine-grained recognition problem without using part and bounding box information and applies this architecture to generic object recognition datasets with superior performance over the directly fine-tuned CNN architecture.
- Rethinking Confidence Calibration for Failure Prediction69
It is revealed that popular confidence calibration methods often lead to worse confidence separation between correct and incorrect samples, making it more difficult to decide whether to trust a prediction or not, and proposes a simple hypothesis: flat minima is beneficial for failure prediction.
- Contextual text/non-text stroke classification in online handwritten notes with conditional random fields69
Experimental results on a publicly available database of freely hand-drawn documents demonstrate the superiority of the approach and the benefit of contextual information combination for solving text/non-text classification in online handwritten documents.
- High accuracy handwritten Chinese character recognition using LDA-based compound distances67
To improve the accuracy of handwritten Chinese character recognition (HCCR), the proposed linear discriminant analysis (LDA)-based compound distances for discriminating similar characters reduces the error rates by factors of over 26% and demonstrates the superiority of LDA-based method over the CMF andThe superiority of discriminant vector learning from high-dimensional feature spaces.
- 66
- 64
- 63
- Image-to-Markup Generation via Paired Adversarial Learning63
Experimental results show that the proposed scalable paired adversarial learning method for image-to-markup generation can significantly improve the generalization performance over standard attention-based encoder-decoder models.
- LightweightNet: Toward fast and lightweight convolutional neural networks via architecture distillation62
The LightweightNet, designed based on a comprehensive analysis of the network architecture, consists of network parameter compression, network structure acceleration, and non-tensor layer improvement and can effectively reduce the computational cost and the model size by > 4 × with negligible accuracy loss.
- 61
- 61
- An approach for real-time recognition of online Chinese handwritten sentences59
A real-time recognition approach for sentence-based input of Chinese handwriting using a dynamically maintained segmentation-recognition candidate lattice that integrates multiple contexts including character classification, linguistic context and geometric context is proposed.
- Classification and Learning Methods for Character Recognition: Advances and Remaining Problems59
This chapter briefly review the learning-based classification methods that have been successfully applied to character recognition, with a special section devoted to the classification of large category set.
- 57
- Cross-Modal Causal Representation Learning for Radiology Report Generation56
A two-stage framework named Cross-Modal Causal Representation Learning (CMCRL), consisting of the Radiological Cross-modal Alignment and Reconstruction Enhanced (RadCARE) pre-training and the Visual-Linguistic Causal Intervention (VLCI) fine-tuning, which establishes robust cross-modal correlations even with incomplete data.
- 55
- Preprocessing and statistical/structural feature extraction for handwritten numeral recognition55
Efficient preprocessing and feature extraction algorithms are presented to achieve high accuracy for handwritten numeral recognition and the efficiency of the proposed method has been demonstrated in recognition experiments on constrained and unconstrained handwritten numerals.
- Realtime multi-scale scene text detection with scale-based region proposal network54
This paper proposes a two-stage framework for realtime multi-scale scene text detection using a novel Scale-based Region Proposal Network (SRPN) which can localize text of wide scale range and estimate text scale efficiently and a Fully Convolutional Network based scene text detector.
- Handwritten Chinese Character Recognition: Effects of Shape Normalization and Feature Extraction54
The technology of handwritten Chinese character recognition (HCCR) has seen significant advances in the last two decades owing to the effectiveness of many techniques, especially those for character shape normalization and feature extraction, but are insufficient applied to unconstrained handwriting.
- Fast instance selection for speeding up support vector machines53
Clustering here is to improve the efficiency of instance selection, other than to select instances directly from clusters as previous methods did, and selects fewer instances than most competitive algorithms and keeps the highest classification accuracy on most datasets.
- Class-specific feature polynomial classifier for pattern classification and its application to handwritten numeral recognition53
The CFPC can be viewed as a hybrid of ordinary PC and projection distance method, and better separate one class from the others, and the incorporation of class-specific projection distance further improves the separability.
- OpenMix: Exploring Outlier Samples for Misclassification Detection52
This work finds that the well-known Outlier Exposure, which is powerful in detecting out-of-distribution samples from unknown classes, does not provide any gain in identifying misclassification errors, and proposes a novel method called OpenMix, which incorporates open-world knowledge by learning to reject uncertain pseudo-samples generated via outlier transformation.
- 51
- 51
- Dewarping Document Image by Displacement Flow Estimation with Fully Convolutional Network50
This paper proposes a novel framework for both rectifying distorted document image and removing background finely, by estimating pixel-wise displacements using a fully convolutional network (FCN).
- Adaptive spatial pooling for image classification50
An adaptive spatial pooling method for enhancing the discriminability of feature representation for image classification by formulating the pooling distribution learning and classifier training jointly, which can extract multiple spatial layouts of arbitrary shapes rather than regular rectangular regions.
- A fast projected fixed-point algorithm for large graph matching50
Results show that the proposed fast algorithm for approximate matching of large graphs is more than an order of magnitude faster than the previous state-of-the-art algorithms, while maintaining comparable accuracy in large graph matching.
- 50
- 48
- Graph-to-Graph: Towards Accurate and Interpretable Online Handwritten Mathematical Expression Recognition47
This paper formulates the HMER as a graph-to-graph (G2G) learning problem, and explicitly resolves the symbol segmentation problem while still being trained end- to-end, making the whole system much more accurate and interpretable.
- Geometry preserving multi-task metric learning47
The novel regularized approach extends previous methods from the vector regularization to a general matrix regularization framework and by exploiting von Neumann divergence as the regularization, the new multi-task metric learning method has the capability to well preserve the data geometry.
- 46
- 46
- 46
- 44
- Multi-view Semi-supervised 3D Whole Brain Segmentation with a Self-ensemble Network43
A semi-supervised segmentation method based on deep neural networks to exploit the plenty of unlabeled data by extending the self-training method, and improve the U-Net model by designing a novel self-ensemble architecture and a random patch-size training strategy to reduce the model storage and computational cost is proposed.
- 43
- 42
- Document Dewarping with Control Points42
A simple yet effective approach to rectify distorted document image by estimating control points and reference points, which can rectify document images with various distortion types, and yield state-of-the-art performance on real-world dataset.
- MuLTReNets: Multilingual text recognition networks for simultaneous script identification and handwriting recognition42
A new framework named multilingual text recognition networks (MuLTReNets) is proposed, which performs superiorly for both script identification and handwriting recognition and outperforms cascade systems but also surpasses systems particularly designed for specific scripts.
- Do acupuncture points exist?42
Three of the four acupuncture points showed significantly elevated concentrations of elements Ca, Fe, Cu and Zn in relation to levels in the surrounding tissue, with similar elevation ratios for Cu and Fe.
- 42
- 41
- 40
- The Formation and Transformation Mechanisms of Deep Consumer Engagement and Purchase Behavior in E-Commerce Live Streaming40
This study constructed a model of the formation and transformation mechanisms of deep consumer engagement and purchase behavior in an e-commerce live streaming environment via the grounded theory coding of interview data from 42 consumers.
- Estimates of the global, regional, and national burden of atrial fibrillation in older adults from 1990 to 2019: insights from the Global Burden of Disease study 201940
From 1990 to 2019, the cases of incidences, deaths, and DALYs have shown a global increase, and the burden of AF increased promptly in the lower SDI regions.
- 40
- 39
- Recognizing online handwritten Chinese characters using RNNs with new computing architectures39
This paper proposes an end-to-end recognizer for online handwritten Chinese characters recognition based on a new recurrent neural network (RNN) that obtains higher recognition accuracy with fewer parameters and two new computing architectures are proposed based on traditional RNN system.
- Multi-task Layout Analysis for Historical Handwritten Documents Using Fully Convolutional Networks39
Experimental results on the public dataset DIVA-HisDB containing challenging medieval manuscripts demonstrate the effectiveness and superiority of the proposed multi-task layout analysis method.
- 39
- 38
- 38
- Discriminative quadratic feature learning for handwritten Chinese character recognition37
By combining dimensionality promotion and reduction, the proposed DQFL method can learn a much more discriminative and nonlinear feature representation, which can then boost the classification accuracy significantly.
- 37
- 36
- 34
- 34
- 33
- 33
- Online semi-supervised learning with learning vector quantization33
This paper proposes a novel OSSL method based on learning vector quantization (LVQ), which exhibits higher accuracy compared with the baseline methods and graph-based methods and is much more efficient than graph- based methods in both training and test time.
- 33
- Error-correcting output codes based ensemble feature extraction33
The decoding strategy using a meta classifier is shown to be more computationally efficient than the linear loss-weighted decoding in state-of-the-art ECOC methods.
- Gaussian Process Latent Random Field32
A novel supervised extension of GPLVM is proposed, called Gaussian process latent random field (GPLRF), by enforcing the latent variables to be a Gaussian Markov random field with respect to a graph constructed from the supervisory information.
- ProtoGCD: Unified and Unbiased Prototype Learning for Generalized Category Discovery31
A dual-level adaptive pseudo-labeling mechanism to mitigate confirmation bias, together with two regularization terms to collectively help learn more suitable representations for GCD are proposed.
- ModalPrompt: Towards Efficient Multimodal Continual Instruction Tuning with Dual-Modality Guided Prompt31
This paper proposes a novel prompt learning framework for MCIT to effectively alleviate forgetting of previous knowledge while managing computational complexity with natural image-text supervision and exploit efficient prompt fusion for knowledge transfer and prompt selection for complexity management with dual-modality guidance.
- Unified Entropy Optimization for Open-Set Test-Time Adaptation31
This paper proposes a simple but effective framework called unified entropy optimization (UniEnt), which is capable of simultaneously adapting to covariate-shifted in-distribution data and detecting covariate-shifted out-of-distribution (csOOD) data.
- Revisiting Confidence Estimation: Towards Reliable Failure Prediction31
This study investigates the problem and reveals that popular calibration and OOD detection methods often lead to worse confidence separation between correctly classified and misclassified examples, making it difficult to decide whether to trust a prediction or not, and proposes to enlarge the confidence gap by finding flat minima.
- 31
- 31
- A robust approach to text line grouping in online handwritten Japanese documents31
An effective approach for grouping text lines in online handwritten Japanese documents by combining temporal and spatial information is presented and a set of performance metrics for evaluating from multiple aspects are provided.
- Aspect Ratio Adaptive Normalization for Handwritten Character Recognition31
Experimental results of multilingual character recognition and numeral recognition demonstrate the advantage of ARAN over conventional normalization method.
- 30
- Anomaly Detection via Minimum Likelihood Generative Adversarial Networks30
A new domain-based anomaly detection method based on generative adversarial networks (GAN) is proposed and minimum likelihood regularization is proposed to make the generator produce more anomalies and prevent it from converging to normal data distribution.
- SDE: A Novel Selective, Discriminative and Equalizing Feature Representation for Visual Recognition30
The SDE representation is a feature learning mechanism by jointly optimizing the pooled representations with the target of learning more selective, discriminative and equalizing features, and uses bilevel optimization to solve the joint optimization problem.
- Traffic Sign Detection Using a Cascade Method With Fast Feature Extraction and Saliency Test30
This paper proposes a fast traffic sign detection method based on a cascade method with saliency test and neighboring scale awareness that achieves competing performance and runs 2~7 times as fast as most of the state-of-the-art methods.
- 30
- Integrating Genomics and Proteomics Data to Predict Drug Effects Using Binary Linear Programming30
A novel approach to infer cell-specific pathways and identify a compounds' effects using gene expression and phosphoproteomics data under treatments with different compounds can be used to elucidate potential mechanisms of a compound's efficacy.
- 30
- PILoRA: Prototype Guided Incremental LoRA for Federated Class-Incremental Learning29
Experimental results on standard datasets indicate that the proposed Incremental LoRA method outperforms the state-of-the-art approaches significantly and exhibits strong robustness and superiority in different settings and degrees of data heterogeneity.
- 29
- Offline signature verification using convolution Siamese network29
The experimental results on two offline public databases demonstrate the superiority of the convolutional Siamese neural network method on the offline signature verification problem.
- 29
- 29
- 29
- 28
- Minimum-risk training for semi-Markov conditional random fields with application to handwritten Chinese/Japanese text recognition28
An alternative parameter learning method by minimizing the risk on the training set, which has unequal misclassification costs depending on the hypothesis and the ground-truth is proposed, which outperforms the best system in competition.
- 28
- 28
- Pattern Field Classification with Style Normalized Transformation27
The Bayes decision theory is extended, and the Field Bayesian Model (FBM) is developed to deal with field classification, to transfer the model to unseen styles and gets significant error rate reduction compared to the state-of-the-art methods.
- 26
- 26
- Aircraft Detection by Deep Convolutional Neural Networks25
Two types of deep learning methods are implemented, deep convolutional neural network (DNN) and deep belief net (DBN), comparing their performances with that of the traditional methods in the task of aircraft detection, and both reduce at least 40% of the false alarm rate of the traditional methods.
- Transcript mapping for handwritten Chinese documents by integrating character recognition model and geometric context25
The experimental results demonstrate the superiority of recognition-based text line alignment and the benefit of integrating geometric context and the tool based on the proposed approach has been practically used for labeling handwritten Chinese documents.
- 25
- An approach for handwritten Chinese text recognition unifying character segmentation and recognition24
- 24
- Fully Convolutional Network Based Skeletonization for Handwritten Chinese Characters24
This paper presents an effective fully convolutional network (FCN) to extract stroke skeletons for handwritten Chinese characters and combines the holistically-nested architecture with regressive dense upsampling convolution (rDUC) and recently proposed hybrid dilated Convolution (HDC) to generate pixel-level prediction for skeleton extraction.
- Text/Non-text Classification in Online Handwritten Documents with Conditional Random Fields24
A Conditional Random Field is utilized for jointly modeling several sources of information that contribute to improve the classification accuracy at the stroke level in unconstrained handwritten online documents.
- 23
- Page Segmentation Using Convolutional Neural Network and Graphical Model23
An effective method for page segmentation using convolutional neural network (CNN) and graphical model, where the CNN is powerful for extracting visual features and the graphical model explores the relationship (spatial context) between visual primitives and regions is proposed.
- 23
- Partial discriminative training for classification of overlapping classes in document analysis23
A partial discriminative training (PDT) scheme is proposed, in which, a training pattern of an overlapping class is used as a positive sample of its labeled class, and neither positive nor negative sample for its allied classes (those overlapping with the labeled class).
- 22
- 22
- 22
- Residual Dual Scale Scene Text Spotting by Fusing Bottom-Up and Top-Down Processing22
This paper proposes a novel text spotter by fusing bottom-up and top-down processing, and proposes a differentiable operator named RoISlide, which can extract features for arbitrary text regions from whole image feature maps and improves the robustness against scale variance.
- 22
- 22
- Confidence Transformation for Combining Classifiers22
The results show that confidence transformation is efficient to improve the combination performance in all the settings, and the confidence transformation methods are highly robust to the validation sample size in parameter estimation.
- Average of Pruning: Improving Performance and Stability of Out-of-Distribution Detection21
An average of pruning (AoP), consisting of model averaging (MA) and pruning, to mitigate the unstable behaviors of OOD detection along the optimization trajectory during training is proposed.
- 21
- 21
- 21
- An over-segmentation method for single-touching Chinese handwriting with learning-based filtering21
An effective over-segmentation method with learning-based filtering using geometric features for single-touching Chinese handwriting that can deal with touching strings with more than two characters.
- 21
- 21
- Efficient Feature Coding Based on Auto-encoder Network for Image Classification20
This paper proposes to use the Auto-Encoder (AE) network as a local descriptor coding block, and further embed AE network in the BoW framework for the purpose of image classification, and adds an efficient and effective regularization term into the learning process of AE network to promote sparsity of the hidden layer for each input descriptor.
- Multi-class segmentation of free-form online documents with tree conditional random fields20
Being fully trainable, the system is shown to properly handle difficult segmentation problems arising in unconstrained online note-taking documents, where no prior knowledge is available regarding the layout or the expected content.
- Evaluation of weighted Fisher criteria for large category dimensionality reduction in application to Chinese handwriting recognition20
The sample level KNN (SKNN) method is shown to outperform significantly other methods in Chinese handwriting recognition such as the locally linear discriminant analysis (LLDA), neighbor class linear discriminatory analysis (NCLDA), and heteroscedastic linear discriminic analysis (HLDA).
- 20
- 20
- Unified Classification and Rejection: A One-versus-all Framework19
This paper forms the open set recognition of K-known-class as a (K+1)-class classification problem with model trained on known-class samples only, and shows that combining the scores of OVA classifiers can give (K+1)-class posterior probabilities, which enables classification and OOD rejection in a unified framework.
- Fuse, Reason and Verify: Geometry Problem Solving with Parsed Clauses from Diagram19
This paper proposes a neural-symbolic model for plane geometry problem solving (PGPS), named PGPSNet-v2, with three key steps: modal fusion, reasoning process and knowledge verification, which outperforms existing symbolic and neural solvers in GPS performance, while maintaining good explainability and reliability.
- 19
- Unsupervised language model adaptation for handwritten Chinese text recognition19
The proposed unsupervised LMA approach improves the recognition performance impressively, particularly for ancient domain documents with the recognition accuracy improved by 7 percent, and the combination of the two compression methods largely reduces the storage size of language models with little loss of recognition accuracy.
- 18
- Decision Controller for Object Tracking With Deep Reinforcement Learning18
A decision controller (DC) which is generally applicable to both SOT and MOT tasks and learns an optimal decision-making policy with a deep reinforcement learning algorithm that maximizes long term tracking performance without supervision is proposed.
- DNA computing inspired deep networks design18
This paper proposes a novel method, called DNA computing inspired networks design (DNAND), to automatically learn high performance deep networks, using DNA strands to represent blocks of a model, and these DNA strands are reacted to construct the overall networks according to the base pairing principle.
- 18
- 18
- A multi-task framework for metric learning with common subspace18
A novel framework to make metric learning benefit from jointly training all tasks is proposed, based on the assumption that discriminative information is retained in a common subspace for all tasks, and can be readily used to extend many current metric learning methods.
- 18
- 18
- 18
- 17
- Learning to Understand Traffic Signs17
This work proposes a novel multi-task learning architecture that extracts text and symbol information from traffic signs, reasons the relationship between texts and symbols, classifies signs into different categories, and finally, composes the descriptions of the signs.
- Contextual Stroke Classification in Online Handwritten Documents with Edge Graph Attention Networks17
The edge graph attention network (EGAT) is proposed to address the stroke classification problem and is formulated as a node classification problem in a relational graph, which is constructed based on the temporal and spatial relationship of strokes.
- Multi-branch guided attention network for irregular text recognition17
The proposed Multi-Branch guided Attention Network (MBAN) makes mutual guidance among multi-branch data in training, so as to learn invariant semantic representation between regular text images and the corresponding irregular images.
- A Sparse Projection and Low-Rank Recovery Framework for Handwriting Representation and Salient Stroke Feature Extraction17
A new Robust Principal Component Analysis (RPCA)-based framework called Sparse Projection and Low-Rank Recovery (SPLRR) is proposed for handwriting representation and salient stroke feature extraction that learns a similarity-preserving sparse projection for extracting salient stroke features and embedding new inputs for classification.
- 17
- 17
- 17
- 17
- PASS++: A Dual Bias Reduction Framework for Non-Exemplar Class-Incremental Learning16
This paper identifies two fundamental causes of forgetting in CIL: representation bias and classifier bias and proposes a simple yet effective dual-bias reduction framework, which leverages self-supervised transformation (SST) in the input space and prototype augmentation (protoAug) in the feature space.
- 16
- 16
- 16
- 16
- 16
- Pattern Recognition16
A novel variational model to jointly estimate geometry and motion from a sequence of light fields captured with a plenoptic camera is presented, which enforces multi-view geometry consistency, and piecewise smoothness assumptions on the scene flow variables.
- Maxi-Min discriminant analysis via online learning16
This work proposes a novel dimensionality reduction method called Maxi-Min Discriminant Analysis (MMDA), which attempts to find a low-dimensional subspace by maximizing the minimal (worst-case) divergence among classes.
- 16
- 16
- 15
- Federated Continual Instruction Tuning15
This work introduces the Federated Continual Instruction Tuning (FCIT) benchmark and proposes a dynamic knowledge organization to effectively integrate updates from different tasks during training and subspace selective activation to allocate task-specific output during inference.
- Social Relation Reasoning Based on Triangular Constraints15
The TRGAT employs the attention mechanism to aggregate features with triangular constraints in the graph, thereby exploiting the higher-order context to reason social relations iteratively, and to acquire better feature representations of persons, node contrastive learning into relation reasoning is introduced.
- 15
- 15
- 15
- Keyword spotting in unconstrained handwritten Chinese documents using contextual word model15
A method for keyword spotting in off-line Chinese handwritten documents using a contextual word model, which measures the similarity between the query word and every candidate word in the document by combining a character classifier and the geometric context as well as linguistic context is proposed.
- Multi-Task Low-Rank Metric Learning Based on Common Subspace15
A novel multi-task metric learning framework based on the assumption that the discriminative information across all the tasks can be retained in a low-dimensional common subspace is proposed which proves effective in suppressing noise and hence more resistant to over-fitting.
- 15
- 15
- 14
- 14
- Joint learning of error-correcting output codes and dichotomizers from data14
This work proposes a novel model for learning the ECOC matrix and dichotomizers jointly from data as a nonlinear programming problem and develops an efficient alternating minimization algorithm to solve it.
- Insect Species Recognition using Sparse Representation14
This paper proposes two class specific sparse representation methods under weakly supervised learning to discriminate insect species which have substantial similarity to each other and shows that the proposed methods perform well in insect species recognition and outperform the state-of-the-art methods on generic image categorization.
- 14
- DCT-DWT-FFT Based Method for Text Detection in Underwater Images13
This paper proposes a novel method for text detection in underwater images through a new enhancement model that performs well for both underwater and natural scene images and outperforms the existing methods on all the datasets.
- Deep Rejoining Model for Oracle Bone Fragment Image13
A deep rejoining model (DRM) is proposed for automatic rejoining of oracle bone fragment images and has stronger adaptability and state-of-the-arts performance in rejoining oracleBone fragment images.
- 13
- Ensemble Quadratic Assignment Network for Graph Matching12
This paper transforms traditional graph matching solvers as single-channel GNNs on the association graph and extend the single-channel architecture to the multi-channel network and proposes a random sampling strategy to reduce the computational complexity and GPU memory usage.
- SignParser: An End-to-End Framework for Traffic Sign Understanding12
This work proposes an end-to-end framework that integrates component detection, content reasoning, and semantic description generation that achieves state-of-the-art performance not only in the final semantic description generation stage but also on multiple subtasks of the CASIA-Tencent CTSU Dataset.
- Bilateral upper extremity trunk model for cross-country sit-skiing double poling propulsion: model development and validation12
A bilateral upper extremity trunk (BUET) musculoskeletal model suitable for cross-country sit-skiing based on OpenSim software and muscle-driven ability was proven and will be available online to study the muscle activation of shoulder joint movement.
- Proxy Graph Matching with Proximal Matching Networks12
A new learning-based matching framework, which is designed to be rotationally invariant, which consists of three parts: a graph neural network to generate a high-level local feature, an attention-based module to normalize the rotational transform, and a global feature matching module based on proximal optimization.
- 12
- 12
- Multivariate frequency analysis of urban rainfall characteristics using three-dimensional copulas12
This study presents a three-dimensional copula-based multivariate frequency analysis of rainfall characteristics based on a long term (1961-2012) rainfall data from Guangzhou, China, and showed that the copulas-based method is easy to implement and can better reflect urban rainstorm characteristics.
- 12
- 12
- The Hitchhiker’s Guide to Autonomous Research: A Survey of Scientific Agents11
This survey bridges the existing gap by presenting a comprehensive blueprint for scientific agents’ design and introduces a unified taxonomy based on capability envelope and capability maturity, characterizing both the scope of scientific workflow coverage and the reliability of agent behavior under realistic research conditions.
- 11
- 11
- 11
- Online Handwritten Diagram Recognition with Graph Attention Networks11
This paper presents a general framework for online handwritten diagram recognition based on graph attention networks (GAT), which model each diagram as a graph in which nodes represent strokes and edges represent the relationships between strokes.
- 11
- 11
- 11
- Special issue on deep learning for document analysis and recognition11
Now researchers are applying deep learning to additional document analysis problems, including layout analysis, writer identification and document retrieval, which has significantly reshaped Document Analysis and Recognition research.
- 11
- Short Text Feature Enrichment Using Link Analysis on Topic-Keyword Graph11
A novel feature enrichment method for short text classification based on the link analysis on topic-keyword graph and the K-L divergence is integrated with the structural similarity to discover the most related keywords.
- Character confidence based on N-best list for keyword spotting in online Chinese handwritten documents11
This paper proposes to directly estimate the posterior probability (also called confidence) of candidate characters based on the N-best paths from the candidate segmentation-recognition lattice using soft-max, and compares the proposed method with some reference methods.
- Graphical lasso quadratic discriminant function and its application to character recognition11
This paper proposes a new classification method called the graphical lasso quadratic discriminant function (GLQDF), which can estimate the covariance matrix (and its inverse) more precisely by exploiting a coordinate descent procedure for the lasso.
- 11
- Vector-to-Image Transformation of Character Patterns for On-line and Off-line Recognition11
The proposed method generates images of various stroke shapes using four painting modes: constant line mode, proportional mode and two calligraphic modes, showing thatcalligraphic painting mode is appropriate for offline recognition while constant line Mode and proportional mode are appropriate for on-line recognition.
- 10
- 10
- 10
- 10
- 10
- Large-Scale Graph-Based Semi-Supervised Learning via Tree Laplacian Solver10
The method consists of two steps: it first approximates a graph by a minimum spanning tree, and then solves the tree-induced quadratic objective function in O(n) time which is the main contribution of this work.
- 10
- 10
- 10
- 10
- 10
- 10
- 9
- 9
- 9
- 9
- 9
- One-Stage Open Set Object Detection with Prototype Learning9
This paper proposes a one-stage open set object detection method based on prototype learning that shows superior performance on identifying objects of both known and unknown classes from images in the open set scenario and inherits all advantages of YOLO v3 such as the high inference speed and the ability of multi-scale detection.
- Handwritten Text Recognition with Convolutional Prototype Network and Most Aligned Frame Based CTC Training9
In the new CTC framework, the blank is formulated as the rejection of character classes and use the one-vs-all prototype classifier as the output layer of the convolutional neural network.
- 9
- 9
- 9
- Fast Genre Classification of Web Images Using Global and Local Features9
A hierarchical algorithm for efficiently classifying Web images into four classes, namely, natural scene images, born-digital images, scanned and cameracaptured paper documents, which are the most prevalent image types on the Web is proposed.
- 9
- 9
- 9
- 9
- 9
- X-ray diffraction-enhanced imaging of uterine leiomyomas.9
X-ray diffraction-enhanced imaging can clearly show internal microstructures of uterine leiomyomas, making the complex procedure of doing this with a large number of pathology slices avoidable.
- 9
- 9
- 8
- 8
- 8
- 8
- 8
- Incremental Learning Vector Quantization for Character Recognition with Local Style Consistency8
The results show that utilizing local style consistency can improve the accuracies of both two test scenarios, and for both supervised and active incremental learning modes.
- 8
- 8
- Transductive Learning on Adaptive Graphs8
By incorporating label information, this work seeks to enhance the performance of graph-based semi-supervised learning by learning the graph and label inference simultaneously by proposing an iterative algorithm to solve the optimization problem which has closed-form solution in each step.
- 7
- 7
- 7
- Upper limb isokinetic muscle strength predicts the performance in cross-country sit-skiing7
Muscle strength and biomechanical analysis of DP process indicated that upper limb extensor muscle strength and muscle coordination were important factors for the power output generation in sit-skiing DP.
- 7
- 7
- 7
- 7
- An evaluation of statistical methods in handwritten hangul recognition7
This study comprehensively evaluate state-of-the-art statistical methods in HHR and implemented fifteen character normalization methods, five feature extraction methods, and four classification methods and evaluated their performances on two public handwritten Hangul databases.
- 7
- Subspace Regularization: A New Semi-supervised Learning Method7
This work introduces into semi-supervised learning the classic low-dimensionality embedding assumption, stating that most geometric information of high dimensional data is embedded in a low dimensional manifold.
- 7
- 7
- 6
- PDE-Agent: A toolchain-augmented multi-agent framework for PDE solving6
PDE-Agent is introduced, the first toolchain-augmented multi-agent collaboration framework, inheriting the reasoning capacity of LLMs and the controllability of external tools and enabling automated PDE solving from natural language descriptions, to validate and evaluate this new paradigm of toolchain-augmented multi-agent PDE solving.
- Air Quality Prediction With a Meteorology-Guided Modality-Decoupled Spatiotemporal Network6
Modality-decoupled spatio-temporal network (MDSTNet) is proposed, an encoder–decoder framework that takes historical air quality observations and multipressure-level meteorological data as input, explicitly modeling them as distinct modalities, while uniquely leveraging multistep weather forecasts as dynamic prompts to guide prediction.
- 6
- 6
- Imitating the oracle: Towards calibrated model for class incremental learning6
This paper thoroughly analyzes the difference between CIL and JT in feature space and weight space and proposes two types of calibration: feature calibration and weight calibration to imitate the oracle (ItO), i.e., JT.
- 6
- Adaptive Scaling for Archival Table Structure Recognition6
An instance segmentation based approach for archival table structure recognition which utilizes both foreground cell content and background ruling line information is proposed which can outperform the baselines and achieve new state-of-the-art performance.
- Insect Recognition Under Natural Scenes Using R-FCN with Anchor Boxes Estimation6
This paper exploits multi-class object detection to eliminate interferences from complex backgrounds, while taking advantages of deep learning to significantly improve the performance of recognition.
- Keyword spotting in handwritten chinese documents using semi-markov conditional random fields6
A document indexing method for keyword spotting based on semi-Markov conditional random fields (semi-CRFs), which provide a theoretical framework for fusing the information of different contexts, can significantly improve the performance compared with the previous character-synchronous dynamic search (CSDS) method.
- 6
- 6
- Context modeling for text/non-text separation in free-form online handwritten documents6
A comprehensive investigation of the sources of contextual information that can benefit the task of discerning textual from non-textual strokes in handwritten online documents and the benefits of combining complementary sources of context for improving the text/non-text recognition performance are presented.
- 6
- 6
- 6
- 6
- 6
- Multiresolution locally expanded HONN for handwritten numeral recognition6
A neural network architecture, multiresolution locally expanded high order neural network (MRLHONN) to solve the problem of handwritten numeral recognition and the recognition results demonstrate the efficiency of the proposed architecture.
- 5
- 5
- 5
- Pulmonary nodule detection in x-ray images by feature augmentation and context aggregation5
PN-DetX incorporates feature fusion and self-attention into x-ray based pulmonary nodule detection tasks, achieving improved detection performance and aspirations for the method to serve as an inspiration for future research in the field of pulmonary nodule detection.
- 5
- 5
- 5
- 5
- 5
- A benchmark for unconstrained online handwritten Uyghur word recognition5
A database of Uyghur online handwritten words is presented and the first benchmark experiments using it are carried out, where the proposed 1D convolutional model and its variations surpassed the base recurrent layered model on the out-of-vocabulary words by clear margin.
- Deep Learning Based Handwritten Chinese Character and Text Recognition5
In HCCR, the traditional normalization-cooperated direction-decomposed feature map is integrated with the deep convolutional neural network, and under this framework, the needs for data augmentation and model ensemble are eliminated.
- Incremental Adaptive Learning Vector Quantization for Character Recognition with Continuous Style Adaptation5
A continuous incremental adaptive learning vector quantization (CIALVQ) method, which incrementally learns a self-adaptive style transfer matrix for mapping input patterns from style-conscious space onto style-free space, which improves the accuracies of both two test scenarios and for both supervised and active incremental learning modes.
- Margin-Aware Binarized Weight Networks for Image Classification5
A margin-aware binarization strategy is constructed by adding a weight constraint into the objective function of deterministic scheme to minimize the margins between real-valued weights and boundaries to learn better binarized weight networks.
- 5
- Learning confidence transformation for handwritten Chinese text recognition5
A string-level confidence-learning method is proposed, which estimates confidence parameters by directly optimizing the performance of character string recognition, and is shown to effectively improve the string recognition performance in experiments of online handwritten Chinese text recognition.
- 5
- Stereo vision system based on the laser dot matrix projection5
A new binocular stereo vision system based on the laser dot matrix projection is presented and a 3D shape measurement of a black cylinder is completed to verify the validity of the system.
- 5
- Radical-Based Hybrid Statistical-Structural Approach for Online Handwritten Chinese Character Recognition5
This chapter describes a new radical-based online handwritten Chinese character recognition approach which combines the advantages of statistical methods andradical-based structural methods and experimental results demonstrate the effectiveness of the proposed approach.
- 5
- 5
- 5
- Grouping Text Lines in Online Handwritten Japanese Documents by Combining Temporal and Spatial Information5
In experiments on the TUAT Kondate database, the proposed approach achieves the Entity Detection Metric (EDM) rate, the Edit-Distance Rate (EDR) of 0.1234, which demonstrates the superiority of the approach.
- 4
- Branch-Tuning: Balancing Stability and Plasticity for Continual Self-Supervised Learning4
This article employs centered kernel alignment (CKA) for quantitatively analyzing model stability and plasticity, revealing the critical roles of batch normalization (BN) layers for stability and convolutional layers for plasticity.
- MeteorPred: A Meteorological Multimodal Large Model and Dataset for Severe Weather Event Prediction4
The first large-scale multimodal dataset for severe weather events prediction, comprising 421,363 pairs of raw multi-year meteorological data and corresponding text caption, and a Meteorology Multimodal Large Model that directly ingests 4D meteorological inputs are introduced.
- 4
- Towards Non-Exemplar Semi-Supervised Class-Incremental Learning4
This work proposes a non-exemplar semi-supervised CIL framework with contrastive learning and semi-supervised incremental prototype classifier (Semi-IPC), enabling the model to incrementally learn from partially labeled new data while maintaining the knowledge of old classes.
- 4
- 4
- 4
- Lateral interaction by Laplacian‐based graph smoothing for deep neural networks4
The authors’ derivation and experiments show that lateral interaction implemented by Linear self‐organising map model is a special case of LS‐regulated k‐means, and they both show the topology‐preserving capability.
- 4
- 4
- 4
- 4
- Rapid Prediction and Evaluation of COVID-19 Epidemic in the United States Based on Feature Selection and Improved ARIMAX Model4
The implementation of the computer program shows that the traditional ARIMAmodel cannot predict the cumulative number of COVID-19 diagnoses in the United States well, and the improved ARIMAX model using features based on stepwise regression can accurately predict the scale of the CO VID-19 epidemic in theUnited States under the 95% confidence interval.
- 4
- 4
- Teaching machines to write like humans using L-attributed grammar4
This paper teaches machines to write like teaching a child by telling the machine the method for writing each character using L-attributed grammar, and demonstrates that the proposed method is effective in generating characters with a complex structure.
- 4
- Discriminative structure learning of sum–product networks for data stream classification4
An online discriminative approach for SPNs for learning both the structure and parameters by estimating the goodness of model fitting of data points and dynamically maintaining a certain amount of informative examples over time, which leads to a fast yet powerful optimization procedure and improved discrimination capability between the genuine class and rival classes.
- Pattern Recognition, Part 1 [Guest editors' introduction]4
This special issue reports the advances in pattern recognition theory and applications, particularly, the basic issues of pattern classification and image analysis and their applications.
- A genome-wide loss-of-function screening method for minimizing false-negatives caused by functional redundancy4
A genome-wide loss-of-function screening method for minimizing false-negatives caused by functional redundancy and its applications in medicine and vaccines.
- 4
- KEYWORD SPOTTING FROM ONLINE CHINESE HANDWRITTEN DOCUMENTS USING ONE-VERSUS-ALL CHARACTER CLASSIFICATION MODEL4
The experimental results on a large handwriting database CASIA-OLHWDB justify the superiority of one-versus-all trained classifiers and the benefits of confidence transformation, character confusion and adaptive thresholds.
- Learning ECOC and Dichotomizers Jointly from Data4
This paper forms a learning model as a sequence of concave-convex programming problems and develops an efficient alternative minimization algorithm to solve it.
- 4
- Polynomial Network Classifier with Discriminative Feature Extraction4
This paper aims to further improve the performance of the subspace-feature-based PNC, and shows that DFE can either improve the classification accuracy or reduce the network complexity.
- 3
- Multimodal Agricultural Agent Architecture (MA3): A New Paradigm for Intelligent Agricultural Decision-Making3
A unified backbone for sugarcane disease classification and detection tools, as well as a sugarcane disease expert model, and an innovative tool selection module are proposed, which develop a multimodal agricultural agent capable of effectively performing tasks in classification, detection, and VQA.
- 3
- 3
- Scene text recognition via dual character counting-aware visual and semantic modeling network3
This work proposes a novel scene text recognizer with a dual character counting-aware visual and semantic modeling network, where the counting information is fused in both vision and language branches.
- 3
- 3
- 3
- SNAP: Shaping neural architectures progressively via information density criterion3
Inspired by the streamline of water drop shaped by air resistance, an information density criterion is defined (play the role of resistance) to drive the network architecture reducing the size of the part with the lowest information density.
- Simulation Prediction and Control Strategy of COVID-19 Dynamic Contact Network in USA and Various States -- Based on Effective Regeneration Number and Improved Discrete SEIQDHR Model3
The classic dynamics model of the spread of infectious diseases, simulates and predicts the future trend of the COVID-19 epidemic in the United States and various states is improved and an improved discrete SEIQDHR model is established for the spread.
- 3
- 3
- Chinese Handwriting Generation by Neural Network Based Style Transformation3
Experimental results on online Chinese handwritten characters show that the proposed method can generate personal-style samples which appear to be naturally written.
- 3
- 3
- Manifold Regularized Multi-Task Learning3
This paper proposes a novel manifold regularized framework for multi-task learning to learn task decision functions as well as a manifold structure from data simultaneously and shows that it can outperform the other competitive methods.
- 3
- 3
- 3
- 3
- 3
- 3
- Partial Discriminative Training of Neural Networks for Classification of Overlapping Classes3
A partial discriminative training (PDT) scheme for neural networks, in which, a training pattern of an overlapping class is used as a positive sample of its labeled class, and neither positive nor negative sample for its allied classes (classes overlapping with the labeled class).
- 3
- 2
- A Lightweight Context-Driven Training-Free Network for Scene Text Segmentation and Recognition2
A novel, training-free plug-and-play framework that leverages the strengths of pre-trained text recognizers while minimizing redundant computations, and introduces an attention-based segmentation stage, which refines candidate text regions at the pixel level, improving downstream recognition.
- 2
- 2
- 2
- 2
- 2
- Multi-scale Unified Network for Image Classification2
A Multi-scale Unified Network (MUSN) consisting of multi-scale subnets, a unified network, and scale-invariant constraint to maintain feature consistency across different scales is proposed.
- 2
- 2
- 2
- 2
- 2
- Editorial: Special Issue on Recent Advances in Cognitive Learning and Data Analysis2
The first selected paper “Action Recognition with a MultiView Temporal Attention Network” introduces global temporal attention pooling and feature-level multi-view fusion as part of a novel action recognition model based on a multiview temporal attention mechanism that can improve the performance of action recognition.
- Adaptable Global Network for Whole-Brain Segmentation with Symmetry Consistency Loss2
This paper proposes a memory-efficient framework to exploit the global brain structure for whole-brain segmentation, and proposes a semi-supervised method based on a symmetry consistency loss and a prior knowledge-based pseudolabel generation strategy.
- 2
- 2
- 2
- Table detection and cell segmentation in online handwritten documents with graph attention networks2
A multi-task learning approach for table detection and cell segmentation with densely connected graph attention networks in free form online documents using a graph attention network model to classify nodes and edges simultaneously.
- 2
- Editorial for special issue on “Advanced Topics in Document Analysis and Recognition”2
This special issue is aimed to report the latest advances of document analysis and recognition on complex documents using advanced techniques and can be divided into two categories, with the contributions outlined below.
- Research on Monitoring Methods for Electricity Hall Staff Based on Autonomous Updating and Semi-supervising Model2
A self-updating and semi-supervised model used to optimize the existing monitoring system and can meet the requirements of high timeliness for personnel on-the-job monitoring, and can also identify the appointed staff on the basis of facial features mining data in batches.
- BYY harmony learning of t-mixtures with the application to image segmentation based on contourlet texture features2
This paper proposes a gradient BYY harmony learning algorithm that can automatically determine the number of actual t-distributions in a dataset during parameter learning and demonstrates by simulation experiments that this proposed algorithm for t-mixtures is both effective and stable on model selection and parameter estimation.
- Automatic Microblog Summarization Based on Unsupervised Key-Bigram Extraction2
An automatic microblog summarization method based on unsupervised key-bigram extraction that was shown to perform superiorly in experiments on SinaWeibo and Twitter datasets is proposed.
- 2
- An Algorithm for Maximum Common Subgraph of Planar Triangulation Graphs2
This paper focuses on a special class of graphs, i.e. Planar Triangulation Graphs, which are commonly used in computer vision, pattern recognition and graphics and proposes a new fast algorithm for solving the Maximum Common Subgraph (MCS) problem.
- An Effective Character Separation Method for Online Cursive Uyghur Handwriting2
This work proposes a character separation method for over-segmentation in online cursive Uyghur handwriting and demonstrates that the proposed method can give a high recall rate of segmentation point detection.
- 2
- 2
- 2
- 2
- 2
- 2
- A pooled subspace mixture density model for pattern classification in high-dimensional spaces2
This paper proposes a new mixture density model based on pooled subspace, where the Gaussian components of each class share a subspace and the complementary subspace is incorporated in the density function.
- Clinical study on treatment of non-small cell lung cancer by chinese herbal medicine combined with bronchial arterial chemotherapy2
Therapeutic effect of BAC could be enhanced by combining it with CHM, and the improvement of clinical symptoms, Karnofsky scoring, body weight and peripheral blood figure in Group A was superior to those in Group B.
- 2
- 1
- 1
- 1
- 1
- Topologically Consistent Prototype Network for Incomplete Multimodal Learning1
Experimental results on multiple multimodal datasets demonstrate the superiority of the proposed framework over existing methods, particularly in small sample learning and imbalanced learning scenarios.
- 1
- 1
- 1
- 1
- 1
- 1
- Three new species of oribatid mites (Acari: Oribatida) from an apple orchard in Beijing, China1
Three new species of oribatid mites are described and illustrated, based on materials collected from an apple orchard in Beijing, China.
- 1
- 1
- 1
- 1
- 1
- 1
- Accelerating Bag-of-Words with SOM1
A fast Bag-of-Words (BoW) method for image classification, inspired by the mechanism that arrangement of neurons in visual cortex can preserve the topology of mapping from inputs, and the fact that human brain can retrieve information almost instantly.
- 1
- A Flattened Maximally Stable Extremal Region Method for Scene Text Detection1
An efficient method, called flattening method, to quickly prune the large number of overlapping MSERs, so as to improve the speed and accuracy of MSER-based scene text detection.
- 1
- 1
- 1
- Pattern Recognition, Part 21
The selected articles address visual categorization by cross-domain dictionary learning, facial expression recognition, face sketch-photo matching, heartbeat rate measurement from facial video, driver gaze estimation, nighttime vehicle detection, and overlaid arrow detection in biomedical images, respectively.
- 1
- 1
- 1
- Insect Recognition Using Sparse Coding and Decision Fusion1
Experimental results on a Tephritidae dataset show that the three proposed discriminative coding methods perform well in insect recognition, and the proposed fusion scheme improves the recognition accuracy significantly.
- Automatic Non-negative Matrix Factorization Clustering with Competitive Sparseness Constraints1
A novel efficient algorithm called the automatic NMF clustering method with competitive sparseness constraints (autoNMF) which can perform the reasonable clustering without pre-assigning the exact number of clusters is proposed.
- 1
- 1
- 1
- 1
- MODIS Image Fusion for Land Cover Classification1
The SFIM transform with b2 can be used to improve the spatial resolution and accuracy of MODIS large areas land cover classification and can improve classification accuracy to a large extent with relative smaller spectra distortion.
- Multiple Outlooks Learning with Support Vector Machines1
A novel MOL framework capable of learning jointly the mapping functions and the classifier in the common latent space is proposed and a new model called MOL-SVM only needs to solve a sequence of standard linear SVM problems and converges rather rapidly within only a few steps is proposed.
- 1
- 1
- Handwritten Numeral String Recognition : Effects of Character Normalization and Feature Extraction(String Recognition, Document Image Understanding and Digital Documents)1
The superiority of moment normalization with adaptive aspect ratio mapping and the gradient direction feature is shown, and it is observed that slant correction is beneficial to string recognition when combined with good normalization methods.
- Handwritten Numeral String Recognition: Effects of Character Normalization and Feature Extraction1
The superiority of moment normalization with adaptive aspect ratio mapping and the gradient direction feature is shown, and it is observed that slant correction is beneficial to string recognition when combined with good normalization methods.
- 1
- –
- SymDiag: Explainable Diagnosis for LLM Reasoning via Neuro-Symbolic Verification–
SymDiag is proposed, a neuro-symbolic framework that reframes reasoning verification as structured failure diagnosis and incorporates a Self-Auditor that disentangles TranslationError from ReasoningError via dual symbolic encodings consistency checks, enabling robust diagnosis under partial observability.
- –
- Two new species of the subgenus Indoribates (Haplozetes) (Acari, Oribatida, Haplozetidae) from Tajikistan–
Two new species of oribatid mites of the subgenus Indoribates (Haplozetes) (Oribatida, Haplozetes) are described based on adult specimens from Tajikistan based on adult specimens recorded for the first time.
- –
- –
- –
- –
- –
- Differentiable Proximal Graph Matching–
This paper relax and decompose the quadratic assignment problem for the graph matching into a sequence of convex optimization problems, and introduces an algorithm for graph matching based on the proximal operator, referred to as differentiable proximal graph matching (DPGM).
- –
- StylePrompter: Enhancing Domain Generalization with Test-Time Style Priors–
This paper proposes to overcome the challenge in previous DG methods by introducing the style prompt in the language modality to adapt the trained model dynamically, and trains a style prompter to extract style information of the current image into an embedding in the token embedding space.
- Towards Reliable Domain Generalization: A New Dataset and Evaluations–
This work proposes a new domain generalization task for handwritten Chinese character recognition (HCCR) to enrich the application scenarios of DG method research and advocates that researchers in the DG community refer to dynamic performance of methods for more comprehensive and reliable evaluation.
- –
- –
- –
- –
- –
- –
- –
- Dynamics analysis and countermeasures of covid-19 epidemic–
The theoretical analysis of the SEIR epidemic model of the new coronavirus shows that prevention and control isolation and medical tracking isolation play an important role in inhibiting the large-scale spread of the epidemic.
- –
- –
- New findings of the subgenus Bischeloribates of Scheloribates (Acari, Oribatida, Scheloribatidae) from China–
Two new species of the subgenus Scheloribates (Bischel oribates) from China are described and illustrated and an identification key to world species of Schelorbates ( BischelorIBates) is given.
- –
- Progressive Scale Expansion Network with Octave Convolution for Arbitrary Shape Scene Text Detection–
Octave Convolution is introduced into scene text detection for enlarging the receptive fields and reducing the spatial redundancy of networks and a state-of-the-art arbitrary shape text detector PSENet is used, which predicts different scale of kernels for each text instance.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Unsupervised Image Segmentation Based on Contourlet Texture Features and BYY Harmony Learning of t-Mixtures–
A novel texture feature extraction mechanism through the contourlet subband coefficient clustering, which is more effective in image segmentation than the discrete cosine transform based normalization technique (DCT), is employed.
- “Special Issue on Handwriting recognition and other PR applications”–
This volume is intended to give the new researchers a good introduction to some of the most promising state-of-the-art researches in the field of Handwriting Recognition and other Pattern Recognition Applications while for an expert, this would serve as a good reference and a source of significant new ideas in this field.
- –
- –
- Graphical Lasso Quadratic Discriminant Function for Character Recognition–
This paper proposes a new classification method called the Graphical Lasso Quadratic Discriminant Function (GLQDF), which can estimate the covariance matrix (and its inverse) more precisely by exploiting a coordinate descent procedure for the lasso.
- Discriminative Training of Subspace Gaussian Mixture Model for Pattern Classification–
This paper proposes a classification method using subspace GMM density model and discriminative training for pattern classification and shows superior classification performance on the MNIST handwritten digit data and UCI datasets.
- Design of Off-Axis TMA System with ZPL in ZEMAX–
This paper introduced a ZPL macros to describe and prevent obscure in Off-axis TMA, an example to show how to use it is also given.
- –
- –
- –
- –
- Analysis of trace elements in breast tumor tissues–
There exist the same kinds of trace elements in various breast tissues,but their relative contents are different, and correlation is positive for Fe and Zn in cancerous tissues, and negative in benign tissues.
- –
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.