Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks151 from public data
- 472
- Look, Imagine and Match: Improving Textual-Visual Cross-Modal Retrieval with Generative Models420
This work proposes to incorporate generative processes into the cross-modal feature embedding, through which it is able to learn not only the global abstract features but also the local grounded features of image-text pairs.
- 322
- Unpaired Image Captioning via Scene Graph Alignments199
This paper proposes an unsupervised feature alignment method that maps the scene graph features from the image to the sentence modality and can generate quite promising results without using any image-caption training pairs, outperforming existing methods by a wide margin.
- 199
- Stack-Captioning: Coarse-to-Fine Learning for Image Captioning194
A coarse-to-fine multi-stage prediction framework for image captioning, composed of multiple decoders each of which operates on the output of the previous stage, producing increasingly refined image descriptions, which simultaneously solves the well-known exposure bias problem and the loss-evaluation mismatch problem.
- 182
- Graininess-Aware Deep Feature Learning for Pedestrian Detection168
This paper proposes a graininess-aware deep feature learning method for pedestrian detection which encodes fine-grained attention masks into convolutional feature maps, which significantly suppresses background interference and highlights pedestrians.
- Learning Common and Specific Features for RGB-D Semantic Segmentation with Deconvolutional Networks166
The proposed feature transformation network achieves competitive segmentation accuracy on NYU depth dataset V1 and V2 and takes advantage of deconvolutional networks which can predict pixel-wise class labels, and develops a new structure for deconvolved of multiple modalities.
- An Empirical Study of Language CNN for Image Captioning159
This paper introduces a language CNN model which is suitable for statistical language modeling tasks and shows competitive performance in image captioning, and is competitive with the state-of-the-art methods.
- Knowledge Distillation via the Target-aware Transformer158
This work proposes a novel one-to-all spatial matching knowledge distillation approach that surpasses the state-of-the-art methods by a significant margin on various computer vision benchmarks, such as ImageNet, Pascal VOC and COCOStuff10k.
- 154
- 145
- EdgeYOLO: An Edge-Real-Time Object Detector128
An efficient, low-complexity, and anchor-free object detector based on the state-of-the-art YOLO framework is proposed, which can be implemented in real time on edge computing platforms and a hybrid random loss function is designed to improve the detection accuracy of small objects.
- Learning Contextual Dependence With Convolutional Hierarchical Recurrent Neural Networks124
This paper integrates CNNs with HRNNs, and develops end-to-end convolutional hierarchical RNNs (C-HRNNs) for image classification, which not only utilize the discriminative representation power of CNNs, but also utilize the contextual dependence learning ability of the authors' HRnns.
- 102
- Learning Discriminative and Shareable Features for Scene Classification94
The experimental results on three challenging scene image classification datasets indicate that the learned filters can achieve very promising performance, and show great complementary effect to the state-of-the-art ConvNets feature.
- Molecular cloning and identification of a flavanone 3-hydroxylase gene from Lycium chinense, and its overexpression enhances drought stress in tobacco85
It is suggested that LcF3H plays a role in enhancing drought tolerance in L. chinense, and its overexpression increases tolerance to drought stress by improving the antioxidant system in tobacco.
- Online latent semantic hashing for cross-media retrieval81
This paper proposes an online optimization scheme towards the challenging task of learning hash functions efficiently on streaming data points, and the computational complexity and memory cost are much less than the size of training dataset at each round.
- Over-Expression of LcPDS, LcZDS, and LcCRTISO, Genes From Wolfberry for Carotenoid Biosynthesis, Enhanced Carotenoid Accumulation, and Salt Tolerance in Tobacco69
This study indicated that LcPDS, LcZDS, and LcCRTISO have the potential to improve carotenoid content and salt tolerance in higher plant breeding.
- Exemplar based Deep Discriminative and Shareable Feature Learning for scene image classification69
A new local feature learning approach named Deep Discriminative and Shareable Feature Learning (DDSFL), which aims to hierarchically learn feature transformation filter banks to transform raw pixel image patches to features and shows great complementary effect to the state-of-the-art Caffe features.
- The effect of pds gene silencing on chloroplast pigment composition, thylakoid membrane structure and photosynthesis efficiency in tobacco plants69
Observations suggest the exact matching of pigments and proteins is important for energy absorption and transfer in photosynthesis and mismatching of them could cause malfunction of PSII.
- Phytoremediation of DEHP and heavy metals co-contaminated soil by rice assisted with a PGPR consortium: Insights into the regulation of ion homeostasis, improvement of photosynthesis and enrichment of beneficial bacteria in rhizosphere soil68
The combined use of rice and the PGPR consortium could remedy DEHP and heavy metal co-contaminated soil economically and ecologically without simultaneously posing risks for rice consumption.
- 59
- 58
- 53
- Overexpression of LcSABP, an Orthologous Gene for Salicylic Acid Binding Protein 2, Enhances Drought Stress Tolerance in Transgenic Tobacco52
The results demonstrate that LcSABP plays a positive regulatory role in drought stress response by enhancing the endogenous SA content, promoting the scavenging of ROS, and regulating of the expression of stress-related transcription factor genes.
- Learning fine-grained features via a CNN Tree for Large-scale Classification52
A novel approach to enhance the discriminability of Convolutional Neural Networks by building a tree structure that could progressively learn fine-grained features to distinguish a subset of classes, by learning features only among these classes.
- Phytoremediation potential, antioxidant response, photosynthetic behavior and rhizosphere bacterial community adaptation of tobacco (Nicotiana tabacum L.) in a bisphenol A-contaminated soil45
Tobacco is a competitive plant in phytoremediation of BPA-contaminated soil, though the growth of tobacco could be inhibited at high concentration of Bpa, and tobacco might promote the removal efficiency of B PA by regulating the rhizosphere bacteria communities.
- 43
- A Single Shot Framework with Multi-Scale Feature Fusion for Geospatial Object Detection41
A large-scale remote-sensing dataset for geospatial object detection (RSD-GOD) that consists of 5 different categories with 18,187 annotated images and 40,990 instances is constructed and a single shot detection framework with multi-scale feature fusion is designed.
- 39
- 37
- Positive feedback regulation of a Lycium chinense-derived VDE gene by drought-induced endogenous ABA, and over-expression of this VDE gene improve drought-induced photo-damage in Arabidopsis37
Data support the possible involvement of endogenous ABA in the positive feedback regulation of LcVDE gene expression in L. chinense under drought stress and may be involved in modulating the level of photosynthesis damage caused by drought stress.
- 34
- 34
- On the sampling of web images for learning visual concept classifiers34
The impact of exploiting the tagged images towards concept learning is studied, investigating the issue of how the quality of pseudo training images affects concept detection performance, and a simple approach is proposed, named semantic field, for predicting the relevance between a target concept and the tag list associated with the images.
- 33
- 33
- Incorporating post-click behaviors into a click model33
A novel click model is introduced, called the post-click click model (PCC), which provides an unbiased estimation of document relevance through leveraging both click behaviors on the search page and post- click behaviors beyond the search pages.
- 32
- Changes in the phenolic compound content and antioxidant activity in developmental maize kernels and expression profiles of phenolic biosynthesis-related genes32
The results revealed that developmental stage affected the accumulation of different phenolic compounds, antioxidant activity and related gene expression and the antioxidant activity of maize kernels was dependent on the total phenolic accumulation which was related to the expression profiles of the genes participated in phenolic synthesis.
- The potential roles of carotenoids in enhancing phytoremediation of bisphenol A contaminated soil by promoting plant physiology and modulating rhizobacterial community of tobacco31
The results demonstrated that exogenous applied β-carotene alleviated the toxic effects of BPA exposure on wild-type tobacco plants after being cultivated for 40 d, and chyb gene overexpressing tobacco showed less phytotoxicity exposed to BPA, through enhancing photosynthetic efficiency and reducing reactive oxygen species (ROS) production.
- Stochastic Downsampling for Cost-Adjustable Inference and Improved Regularization in Convolutional Networks29
This work proposes a novel approach for cost-adjustable inference in CNNs - Stochastic Downsampling Point (SDPoint), which is validated through extensive experiments on image classification.
- 27
- 27
- 27
- Improved fuzzy clustering for image segmentation based on a low-rank prior26
An improved fuzzy clustering algorithm utilizing nonlocal self-similarity and a low-rank prior for image segmentation that can improve efficiency greatly and robustness while achieving satisfactory results is proposed.
- 26
- 25
- De novo characterization of the Lycium chinense Mill. leaf transcriptome and analysis of candidate genes involved in carotenoid biosynthesis24
The transcriptome of L. chinense leaf is characterized and several genes involved in carotenoid biosynthesis in the leaf and the green ripening fruit were found and confirmed by qPCR and high performance liquid chromatography (HPLC).
- 23
- Dual Contrastive Network for Few-Shot Remote Sensing Image Scene Classification22
A transfer-based dual contrastive network (DCN) is proposed, which incorporates two auxiliary supervised contrastive learning branches during the training process, which focus on interclass discriminability and intraclass invariance.
- 22
- 21
- 20
- Comparative transcriptome combined with transgenic analysis reveal the involvement of salicylic acid pathway in the response of Nicotiana tabacum to triclosan stress19
Transgenic tobacco overexpressing a homologous of salicylic acid (SA) binding protein gene was used to assess the SA-mediate TCS tolerance in plant, and showed that transgenic plants exhibited enhanced activities of antioxidant enzymes and stronger TCS resistance than wild-type ones, which verify the important role of SA signal pathway in TCS response of tobacco plants.
- UAV Localization in Low-Altitude GNSS-Denied Environments Based on POI and Store Signage Text Matching in UAV Images18
This paper proposes an autonomous UAV localization method for low-altitude urban scenarios based on POI and store signage text matching (LPS) in UAV images that can achieve stable UAV autonomous localization with a positioning error of around 13 m without knowing the exact initial position of the UAV at take-off.
- 18
- Episodic CAMN: Contextual Attention-Based Memory Networks with Iterative Feedback for Scene Labeling17
- DelugeNets: Deep Networks with Efficient and Flexible Cross-Layer Information Inflows17
Deluge Networks are deep neural networks which efficiently facilitate massive cross-layer information inflows from preceding layers to succeeding layers, and can propagate information across many layers with greater flexibility and utilize network parameters more effectively compared to ResNets, whilst being more efficient than DenseNets.
- 16
- 16
- 15
- Combined application of a PGPR strain Methylophilus sp. FP-6 and adenine to enhance the phytoremediation potential of tobacco: oxybenzone remediation, plant response and rhizosphere bacterial community adaptation14
The findings of this study provide an economical and ecological phytoremediation strategy for the efficient remediation of oxybenzone -contaminated soils.
- Bio-Fabrication of Patterned Cellulose Nano-Fibers14
A new approach for controllable bio-fabrication of patterened cellulose nano-fibers has been proposed by micro-fluidic techniques, via the combination of biological technology and nanotechnology.
- Exploring knowledge of sub-domain in a multi-resolution bootstrapping framework for concept detection in news video13
A model based on a multi-resolution, multi-source and multi-modal (M3) bootstrapping framework that exploits knowledge of sub-domains for concept detection in news video and proposes a "pseudo-Vapnik combined error bound" to tackle the problem of imbalanced distribution of training data in certain segments of sub thedomains.
- 12
- 11
- A Multi-strategy Region Proposal Network11
The novel architecture for region proposal generation is presented which is named as Multi-strategy Region Proposal Network (MSRPN) and can provide important support for the intelligent object detection systems.
- 11
- GRASP: GRAph-Structured Pyramidal Whole Slide Image Representation10
GRASP, a novel lightweight graph-structured multi-magnification framework for processing WSIs in digital pathology, which introduces a convergence-based node aggregation mechanism replacing traditional pooling mechanisms, outperforms state-of-the-art methods by a high margin in terms of balanced accuracy, while being significantly smaller than the closest-performing state-of-the-art models in terms of the number of parameters.
- 9
- Joint-Rerank: a novel method for image search reranking9
A novel image reranking framework—Joint-Rerank is presented which considers multiple modalities of images (or multiple cues) jointly as interdependent attributes of an image entity and two methods are proposed which use different approaches to measure the inter-modal similarities between two nodes.
- Molecular cloning and characterization of a novel carotenoid cleavage dioxygenase 1 from Lycium chinense9
Results indicate that LmCCD1 a novel carotenoids cleavage dioxygenase gene that may regulate the metabolic pathways responsible for aroma metabolite production (such as β‐ionone and pseudoionone) in L. chinense has been identified.
- 8
- 8
- 7
- 7
- Motion feature extraction using magnocellular-inspired spiking neural networks for drone detection6
Experimental results in terms of drone detection performance indicate that the incorporation of MG-SNN significantly improves the accuracy of low-altitude drone detection tasks compared to popular small object detection algorithms, acting as a cheap plug-and-play module in detecting small flying targets against complex backgrounds.
- A Real-Coding Population-Based Incremental Learning Evolutionary Algorithm for Multi-Satellite Scheduling6
A mathematical model of satellite scheduling with complex constraints of load and platform operation is established and a real-coding Population Incremental Based Learning (PBIL) algorithm is proposed to solve the multi-satellite scheduling problem.
- 6
- 5
- A comprehensive swarming intelligent method for optimizing deep learning-based object detection by unmanned ground vehicles5
A boosted RPN (BRPN) with three improvements is developed to solve the problem of restricted exploration space of the region proposal network (RPN), and excellent experimental results are obtained by the deep learning-based object detection method.
- 5
- 5
- 5
- Efficient Perceptual Region Detector Based on Object Boundary5
This work proposes a novel method called Contour-Aware Regions detector (CAR), Inspired by the recent research conclusion of general object proposal methods that contour is important in object localization, to alleviate the problem of super pixle overlapping multi-object regions.
- 4
- Singular value decomposition of split quaternion matrices and its applications for color image processing4
A theory of singular value decomposition of split quaternion matrices using a special 2×2 real representation matrix structure is proposed to provide an effective and efficient algorithm to deal with the color image compression problem.
- 4
- 4
- High resolution remote sensing image ship target detection technology based on deep learning4
The simulation results show that the detection model trained in this paper has the highest recall rate of 98.01% and false alarm rate of 0.83%.
- 4
- Joint-rerank4
Joint-Rerank is presented, which considers image multiple modalities (or multiple cues) jointly as interdependent attributes of an image entity and the relevance scores of multiple modalities are fused to rank the images.
- 4
- 4
- SFPNet: Self-Learning Small Object Detection for Large-Scale Remote Sensing Images3
A self-learning feature point network (SFPNet) is proposed for small object detection, which can captures the geometric information of the object in arbitrary directions as well as enhance the representation of small objects by leveraging feature points.
- AEM-D3QN: A Graph-Based Deep Reinforcement Learning Framework for Dynamic Earth Observation Satellite Mission Planning3
A novel intelligent task scheduling framework that integrates Graph Neural Networks (GNNs) with an Adaptive Exploration Mechanism-enabled Double Dueling Deep Q-Network (D3QN) and offers a scalable and robust solution for real-time satellite mission planning in complex and dynamic operational environments.
- Bidirectional dynamic threshold SNN for enhanced object detection with rich spike information3
BD-SNN is proposed, a new directly trained SNN equipped with Bidirectional Dynamic Threshold neurons (BD-LIF), enhancing the network's information encoding capacity and activation efficiency and validating BD-SNN's superior performance across diverse input scenarios.
- Identification and localization of transmission lines for live working in distribution network3
Three-dimensional LIDAR is selected to design a set of support vector machine (SVM) classifier and correlation analysis based on the target recognition method, which significantly improves the overall operational efficiency of power distribution tasks.
- 3
- Asymmetric similarity-preserving discrete hashing for image retrieval3
A novel Asymmetric Similarity-Preserving Discrete Hashing method to learn compact binary codes for image retrieval by approximating the pairwise similarity matrix in the asymmetric learning manner with two different real-valued embeddings.
- 3
- 3
- Click-through-based Deep Visual-Semantic Embedding for Image Search3
A novel click-through-based deep visual-semantic embedding (C-DVSE) model for learning query and image similarity that is shown to be powerful for keyword-based image search with superior performance over several state-of-the-art methods.
- Target-Aware Fusion: A Diffusion Model for Infrared and Visible Image Integration to Enhance Object Detection2
This article proposes a target detection method based on the visible light infrared fusion diffusion model that introduces the Stable Diffusion architecture and designs a target perception spatial fusion weight module that can adaptively generate a spatial fusion weight map based on modal differences.
- Improving the underwater object detection algorithm for YOLOv52
An improved YOLOv5 algorithm for underwater target detection is proposed to enhance performance in underwater environments that incorporates the Coordinate Attention module to further enhance feature extraction and a BiFPN is employed to better integrate features of different scales.
- Collaborative Global-Local Networks for Fine-Grained Few-shot Learning2
This paper proposes a Collaborative Global-Local Network (CGLNet) to exploit both global and local visual features for few-shot fine-grained classification and shows that the proposed model achieves superior performance to state-of-the-art methods.
- Learning to Prune: General and Efficient Approximate Nearest Neighbor Search with Direction Navigating Graph2
This paper proposes the concept of Direction Navigating Graph (DNG) to offer an online pruning strategy combined with a heuristic search algorithm and develops a neural network that project queries and vertices into a low-dimensional direction vector space.
- ShuffleCloudNet: A Lightweight Composite Neural Network-Based Method for Cloud Computation in Remote-Sensing Images2
A lightweight composite neural network model is designed to calculate the cloud amount in high-resolution visible remote-sensing images by training the model using thumbnail images and browsing images in remote-Sensing images, and the calculation accuracy is better than the classic lightweight networks and backbone networks of the best cloud-detection models.
- Dynamic Task Planning for Heterogeneous Platforms via Spatio-Temporal and Capability Dual-Driven Framework1
A novel mission planning method that integrates spatiotemporal segmentation with Deep Reinforcement Learning (DRL) and establishes a multidimensional spatiotemporal decomposition model to break down complex observation scenarios into manageable subtasks, while incorporating a unified accessibility–visibility computation framework.
- 1
- 1
- 1
- 1
- Conformal Test Martingale-Based Change-Point Detection for Geospatial Object Detectors1
This paper tackles the problem of detecting the moment or change-point when the domain of geospatial images changes based on conformal test martingale and proposes a novel transformation approach to the original conformalTestMartingale to make change- point detection more efficient.
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- Infrared–Visible Image Fusion via Cross-Modal Guided Dual-Branch Networks–
A Cross-modal Guided Dual-Branch Network (CGDBN) that combines convolutional neural networks and Transformer architecture is proposed that combines convolutional neural networks and Transformer architecture for low-altitude aerial drone data fusion.
- –
- –
- –
- –
- Analysis of target recognition technology for autonomous underwater robot–
The problems active and passive sonar face when applied to underwater target detection and their solutions are introduced and the corresponding solutions are listed.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- F3N: Full Feature Fusion Network for Object Detection–
F3N achieves the new state of the art result for one-stage object detection and its accuracies are significantly enhanced compared to the current mainstream approaches such as SSD, DSSD, FPN, YOLO.
- Cross-Modal Multimedia Archives Information Retrieval Based on Semantic Matching–
This paper proposes a cross-modal multimedia archives information retrieval model, carries out feature processing and semantic mapping of archives information, applies the integrated learning method to cross-media archives retrieval, proposes Bagging-SM method for semantic matching of different modes of multimedia archivesInformation retrieval.
- –
- Cloning and characterization of a Lycium chinense carotenoid isomerase gene enhancing carotenoid accumulation in transgenic tobacco–
The results obtained here clearly suggested that the LcCRTISO gene was a promising candidate for carotenoid production.
- Maximally Visual-Homogeneous Region Detector for Large Scale Image Retrieval–
Experiments demonstrate the proposed Maximally Visual-Homogeneous Region (MVHR) detector can find less but representative regions with high repeatability, while still perserving competitive precision compared to the state-of-art detectors for large scale image retrieval.
- –
- –
- Co-optimization of multiple relevance metrics in web search–
This paper proposes two novel relevance optimization approaches, which take different metrics into a global consideration where the objective is to achieve an ideal tradeoff between different metrics.
- Workpiece Recognition by the Combination of Multiple Simplified Fuzzy ARTMAP–
The experimental results have shown that the fusion approach can achieve reliable recognition and the sum rule voting algorithm combines the results from several SFAM's and generates reliable and accurate recognition conclusion.
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 2026-10-12. 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.