Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks206 from public data
- 828
- DER: Dynamically Expandable Representation for Class Incremental Learning707
This work proposes a novel two-stage learning approach that utilizes a dynamically expandable representation for more effective incremental concept modeling and introduces an auxiliary loss to encourage the model to learn diverse and discriminate features for novel concepts.
- Shape-Aware Semi-supervised 3D Semantic Segmentation for Medical Images527
This work proposes a novel shapeaware semi-supervised segmentation strategy to leverage abundant unlabeled data and to enforce a geometric shape constraint on the segmentation output and develops a multi-task deep network that jointly predicts semantic segmentation and signed distance map(SDM) of object surfaces.
- Part-Aware Prototype Network for Few-Shot Semantic Segmentation424
A novel few-shot semantic segmentation framework based on the prototype representation, capable of capturing diverse and fine-grained object features, and a novel graph neural network model to generate and enhance the proposed part-aware prototypes based on labeled and unlabeled images.
- 382
- Distribution Alignment: A Unified Framework for Long-tail Visual Recognition376
This work develops an adaptive calibration function that enables us to adjust the classification scores for each data point in the two-stage learning to balance the class prior, which provides a flexible and unified solution to diverse scenarios in visual recognition tasks.
- SPHINX: The Joint Mixing of Weights, Tasks, and Visual Embeddings for Multi-modal Large Language Models313
SPHINX, a versatile multi-modal large language model (MLLM) with a joint mixing of model weights, tuning tasks, and visual embeddings is presented, and an efficient strategy aiming to better capture fine-grained appearances of high-resolution images is proposed.
- Bipartite Graph Network with Adaptive Message Passing for Unbiased Scene Graph Generation282
This work introduces a novel confidence-aware bipartite graph neural network with adaptive message propagation mechanism for unbiased scene graph generation and proposes an efficient bi-level data resampling strategy to alleviate the imbalanced data distribution problem in training the authors' graph network.
- SentiCap: Generating Image Descriptions with Sentiments248
This work designs a system to describe an image with emotions, and presents a model that automatically generates captions with positive or negative sentiments, and proposes a novel switching recurrent neural network with word-level regularization, able to produce emotional image captions using only 2000+ training sentences containing sentiments.
- 241
- Pose-Aware Multi-Level Feature Network for Human Object Interaction Detection239
This work develops a multi-branch deep network to learn a pose-augmented relation representation at three semantic levels, incorporating interaction context, object features and detailed semantic part cues, and demonstrates its efficacy in handling complex scenes.
- GNeRF: GAN-based Neural Radiance Field without Posed Camera234
GNeRF, a framework to marry Generative Adversarial Networks (GAN) with Neural Radiance Field reconstruction for the complex scenarios with unknown and even randomly initialized camera poses, is introduced and outperforms the baselines favorably in those scenes with repeated patterns or low textures that are regarded as extremely challenging before.
- Learning and Incorporating Top-Down Cues in Image Segmentation218
This paper proposes an approach to utilizing category-based information in segmentation, through a formulation as an image labelling problem, that exploits bottom-up image cues to create an over-segmented representation of an image.
- SGTR: End-to-end Scene Graph Generation with Transformer193
A transformer-based end-to-end framework that first generates the entity and predicate proposal set, followed by inferring directed edges to form the relation triplets, and a new entity-aware predicate representation based on a structural predicate generator that leverages the compositional property of relationships is developed.
- 161
- SemStyle: Learning to Generate Stylised Image Captions Using Unaligned Text125
A model that learns to generate visually relevant styled captions from a large corpus of styled text without aligned images, and a unified language model that decodes sentences with diverse word choices and syntax for different styles.
- 119
- 116
- 108
- HOICLIP: Efficient Knowledge Transfer for HOI Detection with Vision-Language Models105
A novel HOI detection framework that efficiently extracts prior knowledge from CLIP and achieves better generalization is proposed and a training-free enhancement to exploit global HOI predictions from ClIP is proposed.
- Learning Cross-Modal Context Graph for Visual Grounding100
A language-guided graph representation is proposed to capture the global context of grounding entities and their relations, and a cross-modal graph matching strategy for the multiple-phrase visual grounding task is developed.
- From AI for Science to Agentic Science: A Survey on Autonomous Scientific Discovery98
This work establishes a domain-oriented synthesis of autonomous scientific discovery and positions Agentic Science as a structured paradigm for advancing AI-driven research.
- Deep photonic reservoir computing recurrent network98
It is found that the deep PRC owns strong ability to compensate the nonlinearity of fibers, and is applied in the real-world signal equalization of an optical fiber communication system.
- Dynamic Context Correspondence Network for Semantic Alignment98
This paper proposes a context-aware semantic representation that incorporates spatial layout for robust matching against local ambiguities and develops a novel dynamic fusion strategy based on attention mechanism to weave the advantages of both local and context features by integrating semantic cues from multiple scales.
- Learning Multi-Granular Spatio-Temporal Graph Network for Skeleton-based Action Recognition94
A novel multi-granular spatio-temporal graph network is proposed for skeleton-based action classification that jointly models the coarse- and fine-grained skeleton motion patterns and utilises a cross-head communication strategy to mutually enhance the representations of both heads.
- From Pixels to Graphs: Open-Vocabulary Scene Graph Generation with Vision-Language Models81
A new open-vocabulary SGG framework based on sequence generation that leverages vision-language pre-trained models (VLM) by incorporating an image-to-graph generation paradigm and seamlessly integrate explicit relational modeling for enhancing the VL tasks.
- RealDex: Towards Human-like Grasping for Robotic Dexterous Hand79
This paper introduces RealDex, a pioneering dataset capturing authentic dexterous hand grasping motions infused with human behavioral patterns, enriched by multi-view and multimodal visual data, and introduces a cutting-edge dexterous grasping motion generation framework, which aligns with human experience and enhances real-world applicability through effectively utilizing Multimodal Large Language Models.
- 75
- Geometry-Aware Deep Network for Single-Image Novel View Synthesis71
A new region-aware geometric transform network is developed that performs these multiple tasks in a common framework and demonstrates the effectiveness of the network in generating high-quality synthetic views that respect the scene geometry, thus outperforming the state-of-the-art methods.
- DeepPhospho accelerates DIA phosphoproteome profiling through in silico library generation70
A DIA workflow based on in silico spectral libraries generated by a novel deep neural network to expand phosphoproteome coverage is established, which leads to the discovery of more signaling pathways and regulated kinases in an EGF signaling study than the DDA library-based approach.
- A Dual Attention Network with Semantic Embedding for Few-Shot Learning70
This work proposes a novel meta-learning method for few-shot classification based on two simple attention mechanisms: one is a spatial attention to localize relevant object regions and the other is a task attention to select similar training data for label prediction.
- Learning deep structured network for weakly supervised change detection70
This paper presents a weakly supervised approach that needs only image-level labels to simultaneously detect and localize changes in a pair of images and defines a CRF model to incorporate the local differences and context with the dense connections between individual pixels.
- Relation-aware Instance Refinement for Weakly Supervised Visual Grounding69
This paper proposes a novel context-aware weakly-supervised learning method that incorporates coarse-to-fine object refinement and entity relation modeling into a two-stage deep network, capable of producing more accurate object representation and matching.
- Impacts of impaired face perception on social interactions and quality of life in age-related macular degeneration: A qualitative study and new community resources64
Poor face perception in AMD is an important domain contributing to impaired social interactions and quality of life, and this domain should be directly assessed in quantitativequality of life measures, and in resources designed to improve community understanding.
- 63
- Superpixel Graph Label Transfer with Learned Distance Metric60
This work presents a fast approximate nearest neighbor algorithm for semantic segmentation that builds a graph over superpixels from an annotated set of training images and proposes to learn a distance metric that weights the edges in the graph.
- Learning Implicit Temporal Alignment for Few-shot Video Classification58
This work introduces an implicit temporal alignment for a video pair, capable of estimating the similarity between them in an accurate and robust manner, and designs an effective context encoding module to incorporate spatial and feature channel context, resulting in better modeling of intra-class variations.
- Learning Hybrid Models for Image Annotation with Partially Labeled Data57
A hybrid model framework for utilizing partially labeled data that integrates a generative topic model for image appearance with discriminative label prediction is explored and three alternative formulations for imposing a spatial smoothness prior on the image labels are proposed.
- Occlusion Boundary Detection Using Pseudo-depth56
It is shown that crude scaled estimates of depth can be extracted from motion sequences containing a small number of image frames using standard SVD factorization methods followed by weak smoothing using a Markov Random Field defined over super-pixels.
- 54
- General Incremental Learning with Domain-aware Categorical Representations51
This work introduces a flexible class representation based on the von Mises-Fisher mixture model to capture the intra-class structure, using an expansion-and- reduction strategy to dynamically increase the number of components according to the class complexity.
- 51
- 50
- Class-relation Knowledge Distillation for Novel Class Discovery45
A novel knowledge distillation framework is proposed, which utilizes the authors' class-relation representation to regularize the learning of novel classes and develops a learnable weighting function for the regularization, which adaptively promotes knowledge transfer based on the semantic similarity between the novel and known classes.
- An EM Framework for Online Incremental Learning of Semantic Segmentation45
A unified learning strategy based on the Expectation-Maximization (EM) framework, which integrates an iterative relabeling strategy that fills in the missing labels and a rehearsal-based incremental learning step that balances the stability-plasticity of the model.
- 45
- Dynamic Grained Encoder for Vision Transformers44
A Dynamic Grained Encoder for vision transformers is proposed, which can adaptively assign a suitable number of queries to each spatial region and achieves a fine-grained representation in discriminative regions while keeping high efficiency.
- Superpixel-Guided Iterative Learning from Noisy Labels for Medical Image Segmentation44
A superpixel representation is adopted and a robust iterative learning strategy is developed that combines noise-aware training of segmentation network and noisy label refinement, both guided by the superpixels, to exploit the structural constraints in segmentation labels and effectively mitigate the impact of label noise in learning.
- Asynchronous photonic time-delay reservoir computing39
It is proved that the asynchronous reservoir computing shows highly competitive performance on the prediction task of Santa Fe chaotic time series, in comparison with the synchronous counterparts.
- Human-centric Scene Understanding for 3D Large-scale Scenarios38
This paper presents a large-scale multi-modal dataset for human-centric scene under-standing, dubbed HuCenLife, which is collected in diverse daily-life scenarios with rich and fine-grained annotations and designs novel modules for LiDAR-based segmentation and action recognition.
- 38
- 38
- 36
- 35
- Face identity recognition in simulated prosthetic vision is poorer than previously reported and can be improved by caricaturing35
Results show caricaturing improved identity recognition in memory and/or perception (degree by which two faces look dissimilar) down to a resolution of 32 × 32 with 30% phosphene dropout, imply caricatured may offer benefits for patients at resolutions realistic for some current‐generation or in‐development implants.
- A Novel Unified Conditional Score-based Generative Framework for Multi-modal Medical Image Completion34
The UMM-CSGM employs a novel multi-in multi-out Conditional Score Network (mm-CSN) to learn a comprehensive set of cross-modal conditional distributions via conditional diffusion and reverse generation in the complete modality space, and can fit all possible configurations of missing modalities in a single network.
- KD-VLP: Improving End-to-End Vision-and-Language Pretraining with Object Knowledge Distillation33
This paper proposes an object-aware end-to-end VLP framework, which directly feeds image grid features from CNNs into the Transformer and learns the multi-modal representations jointly, and designs two novel pretext tasks by taking object features and their semantic labels from external detectors as supervision.
- Shape-aware Instance Segmentation.31
This paper introduces a novel object segment representation based on the distance transform of the object masks, and designs an object mask network (OMN) with a new residual-deconvolution architecture that infers such a representation and decodes it into the final binary object mask.
- Weakly Supervised Nuclei Segmentation Via Instance Learning30
A modular deep network with two branches: a semantic proposal network and an instance encoding network, which are trained in a two-stage manner with an instance-sensitive loss to enable more effective subtask learning and to promote instance-aware representation learning.
- Pack and Force Your Memory: Long-form and Consistent Video Generation29
MemoryPack is proposed, a learnable context-retrieval mechanism that leverages both textual and image information as global guidance to jointly model short- and long-term dependencies, achieving minute-level temporal consistency.
- 29
- Picture tags and world knowledge28
A novel network estimation algorithm, Inverse Concept Rank, to infer incomplete tag relationships is proposed, and an algorithm for image annotation that takes into account both image and tag features is designed.
- 27
- 27
- DSGG: Dense Relation Transformer for an End-to-End Scene Graph Generation26
A new Transformer-based method, called DSGG, that views scene graph detection as a direct graph prediction problem based on a unique set of graph-aware queries, and utilizes a strategy for relation distillation, aiming to efficiently learn multiple instances of semantic relationships.
- SPHINX: A Mixer of Weights, Visual Embeddings and Image Scales for Multi-modal Large Language Models26
- Scalable wavelength-multiplexing photonic reservoir computing26
It is proved that the multiplexing PRC exhibits a superior performance on the task of signal equalization in an optical fiber communication link, owing to the rich neuron interconnections both in parallel and in series.
- 26
- 26
- 26
- 25
- Topological map learning from outdoor image sequences25
This work proposes an approach to building topological maps of environments based on image sequences, using manifold constraints to find representative feature prototypes so that images can be related to each other, and thereby to camera poses in the environment.
- Machine Learning on Microstructure–Property Relationship of Lithium-Ion Conducting Oxide Solid Electrolytes23
An algorithmic framework is developed to understand the effect of microstructure on the properties by linking the microscopic morphology images to their ionic conductivities, and can determine the microstructural features of a Li-ion conductor with high ionic conductivity, which can guide the synthesis of highly conductive solid electrolytes.
- Novel Class Discovery for Long-tailed Recognition23
This paper proposes an adaptive self-labeling strategy based on an equiangular prototype representation of classes that infers high-quality pseudo-labels for the novel classes by solving a relaxed optimal transport problem and effectively mitigates the class biases in learning the known and novel classes.
- ROI-Constrained Bidding via Curriculum-Guided Bayesian Reinforcement Learning23
This work develops a Curriculum-Guided Bayesian Reinforcement Learning (CBRL) framework to adaptively control the constraint-objective trade-off in non-stationary ad markets and exploits an indicator-augmented reward function free of extra trade-offs.
- Part-aware Prototypical Graph Network for One-shot Skeleton-based Action Recognition22
This paper proposes a part-aware prototypical representation for one-shot skeleton-based action recognition that captures skeleton motion patterns at two distinctive spatial levels and devise a class-agnostic attention mechanism to highlight important parts for each action class.
- 22
- Learning to refine depth for robust stereo estimation21
This paper proposes a novel method to refine the disparity maps generated by deep regression networks by integrating the initial disparity estimation, the confidence map and the disparity gradients into a continuous Markov Random Field for depth refinement, which is capable of representing rich surface structures.
- Scene understanding by labeling pixels21
Pixels labeled with a scene's semantics and geometry let computers describe what they see.
- Weakly-supervised HOI Detection via Prior-guided Bi-level Representation Learning20
A CLIP-guided HOI representation capable of incorporating the prior knowledge at both image level and HOI instance level is developed, and a self-taught mechanism to prune incorrect human-object associations is adopted, showing the efficacy of theHOI representation.
- Modeling Multimodal Aleatoric Uncertainty in Segmentation with Mixture of Stochastic Experts20
This work proposes a novel mixture of stochastic experts (MoSE) model, where each expert network estimates a distinct mode of the aleatoric uncertainty and a gating network predicts the probabilities of an input image being segmented in those modes and yields an efficient two-level uncertainty representation.
- Caricaturing as a general method to improve poor face recognition: Evidence from low-resolution images, other-race faces, and older adults.20
It is shown that caricaturing faces-that is, exaggerating their appearance away from an average face-can provide a useful applied method for improving face recognition across all circumstances including low-resolution images, other-race faces, and in older adult observers.
- 19
- Object Co-detection via Efficient Inference in a Fully-Connected CRF19
This paper forms object co-detection as inference in a fully-connected CRF whose edges model the similarity between object candidates, and learns a similarity function that allows us to efficiently perform inference in this fully- connected graph, even in the presence of many object candidates.
- 19
- P$^2$OT: Progressive Partial Optimal Transport for Deep Imbalanced Clustering18
This paper proposes a novel pseudo-labeling-based learning framework that formulates pseudo-label generation as a progressive partial optimal transport problem, which progressively transports each sample to imbalanced clusters under prior distribution constraints, thus generating imbalance-aware pseudo-labels and learning from high-confident samples.
- 18
- Building Scene Models by Completing and Hallucinating Depth and Semantics18
This work introduces a fully-automatic method to jointly complete and hallucinate depth and semantics in challenging outdoor scenes using a two-layer model representing both the visible information and the hidden one based on the Mumford-Shah functional.
- 18
- 18
- LLM-HD: Layout Language Model for Hotspot Detection with GDS Semantic Encoding17
This work proposes a new layout hotspot detection paradigm, where hotspots are directly detected on binary database files by exploiting a hierarchical GDS semantic representation scheme and a well-designed pre-trained natural language processing (NLP) model.
- 16
- Learning Dynamic Hierarchical Models for Anytime Scene Labeling16
A dynamic hierarchical model for anytime scene labeling that allows us to achieve flexible trade-offs between efficiency and accuracy in pixel-level prediction, and incorporates the cost of feature computation and model inference, and optimizes the model performance for any given test-time budget.
- 16
- Improving face identity perception in age-related macular degeneration via caricaturing15
It is concluded caricaturing has the potential to help many AMD patients recognise faces and was effective in half the eyes with moderate and severe vision loss.
- 14
- Caricaturing faces to improve identity recognition in low vision simulations: How effective is current-generation automatic assignment of landmark points?14
The 68-point caricatures produced significant improvements in identity discrimination relative to veridical, and were approximately 50% as effective as the hand-assigned 147- point caricatures.
- Caricaturing can improve facial expression recognition in low-resolution images and age-related macular degeneration13
Testing potential benefits of caricaturing (exaggerating how the expression's shape differs from neutral) as an image enhancement procedure targeted at mid- to high-level cortical vision found to emerge when improvement was most needed.
- Tracking Large-Scale Video Remix in Real-World Events13
This work proposes visual memes, or frequently re-posted short video segments, for detecting and monitoring such latent video interactions at scale on YouTube, and augments visual memes with text, via a statistical model of latent topics.
- RULER-Bench: Probing Rule-based Reasoning Abilities of Next-level Video Generation Models for Vision Foundation Intelligence12
RULER-Bench is introduced, a benchmark designed to evaluate the reasoning ability of video generation models from the perspective of cognitive rules, and it is expected that the insight obtained will facilitate further development of reasoning-aware video generation, advancing video generation models toward vision foundation intelligence.
- 12
- Auto-segmentation and time-dependent systematic analysis of mesoscale cellular structure in β-cells during insulin secretion12
This approach combines semantic and first-applied instance segmentation to produce separate organelle masks with high Dice and Recall indexes, followed by analysis of organelle localization based on the radial distribution function.
- 12
- Learning structured prediction models for image labeling12
This thesis develops a discriminative labeling framework based on a Conditional Random Field (CRF), in which multiscale feature functions are proposed to capture the image/label contexts at several spatial scales and proposes a modular CRF model that integrates the bottom-up image cues and top-down categorical information.
- 11
- Learning a Layout Transfer Network for Context Aware Object Detection11
This work presents a context aware object detection method based on a retrieve-and-transform scene layout model that provides consistent performance improvements to the state-of-the-art object detection baselines on a variety of challenging tasks in the traffic surveillance and the autonomous driving domains.
- 11
- 11
- A unified model of short-range and long-range motion perception11
A hierarchical model as a unified framework for modeling both short-range and long-range motion perception, consisting of a data likelihood that proposes multiple motion hypotheses using nonlinear matching and a hierarchical prior that imposes slowness and spatial smoothness constraints on the motion field at multiple scales.
- FastGrasp: Efficient Grasp Synthesis with Diffusion10
A novel diffusion-modelbased approach that generates the grasping pose in a one-stage manner and develops a Latent Diffusion Model with an Adaptation Module for object-conditioned hand pose generation and a contact-aware loss to enforce the physical constraints between hands and objects.
- Data-Driven Discovery of Polar Organic Cocrystals: Integration of Machine Learning and Automated Screening10
A data-driven method that combines machine learning (ML) with high-throughput (HT) automation to speed up the discovery of polar organic cocrystals and demonstrates the potential to advance functional molecular materials and pave the way for next-generation applications using polar organic cocrystals.
- 10
- Learning Autonomous Exploration and Mapping with Semantic Vision10
This work addresses the problem of autonomous exploration and mapping for a mobile robot using visual inputs using a learning-based approach based on semantic interpretation of visual scenes and is believed to be the first algorithm that formulate Exploration and mapping into learning framework.
- 10
- 10
- Automatic spinal curvature measurement on ultrasound spine images using Faster R-CNN9
A fully automatic framework based on Faster R-CNN for detecting vertebral lamina and to measure the fitting spinal curves from the detected lamina pairs to provide accurate and reliable automatic curvature measurements on ultrasound spine images for spine deformities.
- Structural Kernel Learning for Large Scale Multiclass Object Co-detection9
This work forms co-detection as the joint multiclass labeling of object candidates obtained in a class-independent manner, and introduces a structural boosting algorithm that lets us exploits rich, high-dimensional deep network features to learn object similarity within the authors' fully-connected CRF.
- The Robust Semantic Segmentation UNCV2023 Challenge Results8
This paper outlines the winning solutions employed in addressing the MUAD uncertainty quantification challenge held at ICCV 2023, which primarily revolved around enhancing the robustness of semantic segmentation in urban scenes under varying natural adversarial conditions.
- 8
- 8
- 8
- 8
- 8
- 7
- 7
- Weakly Supervised Change Detection in a Pair of Images.7
This paper employs a deep neural network with DAG topology to learn patterns of change from image-level labeled training data, and applies a constrained mean-field algorithm to estimate the pixel-level labels, which enables imposing global constraints on the observed foreground probability mass function.
- 7
- 7
- Probabilistic models of vision and max-margin methods7
This paper shows that by placing bounds on the normalization constant the authors can obtain computationally tractable approximations to probabilistic methods including multi-class max- margin, ordinal regression, max-margin Markov networks and parsers, multiple-instance learning, and latent SVM.
- OccluBEV: Occlusion Aware Spatiotemporal Modeling for Multi-view 3D Object Detection6
A multi-head learning framework which jointly models multiple strongly-correlated tasks in a unified BEV space and alleviates the overfitting problem in BEV feature learning under a single task is designed.
- 3D Box Proposals From a Single Monocular Image of an Indoor Scene6
This paper develops an integrated, fully differentiable framework that inherently predicts a depth map, extracts a 3D volumetric scene representation and generates 3D object proposals from a single monocular RGB image.
- 6
- Using latent Dirichlet allocation to incorporate domain knowledge for topic transition detection6
This paper incorporates relevant written materials, e.g., textbooks for lectures, which convey semantic relationships, in particular domain-specific relationships, between words, and trains latent Dirichlet allocation models on these materials and measures the similarity between slides and transcripts in the acquired hidden-topic space.
- 6
- 5
- 5
- CIBench: Evaluating Your LLMs with a Code Interpreter Plugin5
An interactive evaluation framework, named CIBench, is proposed to comprehensively assess LLMs' ability to utilize code interpreters for data science tasks and provides valuable insights for future LLMs in code interpreter utilization.
- Mutual Information-guided Knowledge Transfer for Novel Class Discovery5
The insight is to utilize mutual information to measure the relation between seen classes and unseen classes in a restricted label space and maximizing mutual information promotes transferring semantic knowledge.
- Single Image 3D Object Estimation with Primitive Graph Networks5
A two-stage graph network for primitive-based 3D object estimation, which consists of a sequential proposal module and a graph reasoning module, capable of taking into account rich geometry and semantic constraints during 3D structure recovery, producing 3D objects with more coherent structure even under challenging viewing conditions.
- 5
- 5
- 4
- Learning by Correction: Efficient Tuning Task for Zero-Shot Generative Vision-Language Reasoning4
This work introduces Image-Conditioned Caption Correction (ICCC), a novel pre-training task designed to enhance VLMs' zero-shot performance without the need for labeled task-aware data, and constructs data samples of ICCC task from image-text datasets with low labeling and computation costs.
- 4
- 4
- Towards Purely Unsupervised Disentanglement of Appearance and Shape for Person Images Generation4
An encoder-decoder-like network to extract both the shape and appearance features from input images at the same time, and train the parameters by three losses: feature adversarial loss, color consistency loss and reconstruction loss, which achieves clean disentanglement.
- Object-Aware Dictionary Learning with Deep Features4
This work presents an object-aware dictionary learning framework that systematically incorporates region proposals and deep features in order to improve the discriminative power of the combined classifier and proposes an iterative solver.
- Learning Hough Transform with Latent Structures for Joint Object Detection and Pose Estimation4
A novel max-margin Hough transform with latent structure for joint object detection and pose estimation and adopts a multiple instance learning approach to handle the lack of correspondence among training instances with noisy bounding-box labels.
- 4
- Face Detection And Tracking In Video To Facilitate Face Recognition In A Visual Prosthesis4
This work is developing a real-time system capable of automatically detecting, tracking and zooming onto faces, and facilitating face and expression recognition with low-resolution retinal implant array.
- Gradient-Map-Guided Adaptive Domain Generalization for Cross Modality MRI Segmentation3
A novel adaptive domain generalization framework is proposed, which integrates a learning-free cross-domain representation based on image gradient maps and a class prior-informed test-time adaptation strategy for mitigating local domain shift.
- Fixed-Price Diffusion Mechanism Design3
This work proposes the very first fixed-price mechanism to incentivize the seller's neighbours to inform their neighbours about the sale and to eventually inform all buyers in the network to improve seller's revenue.
- Confidence-Aware Adversarial Learning for Self-supervised Semantic Matching3
This paper introduces a Confidence-Aware Semantic Matching Network (CAMNet) which instantiates two key ideas of the approach, and proposes to estimate a dense confidence map for a matching prediction through self-supervised learning.
- Relation-aware Hierarchical Prompt for Open-vocabulary Scene Graph Generation2
The Relation-Aware Hierarchical Prompting (RAHP) framework is proposed, which enhances text representation by integrating subject-object and region-specific relation information, and introduces a dynamic selection mechanism within Vision-Language Models (VLMs), which adaptively selects relevant text prompts based on the visual content, reducing noise from irrelevant prompts.
- 2
- 2
- SP$^2$OT: Semantic-Regularized Progressive Partial Optimal Transport for Imbalanced Clustering2
This paper introduces a novel optimal transport-based pseudo-label learning framework that formulates pseudo-label generation as a Semantic-regularized Progressive Partial Optimal Transport (SP$^2$OT) problem, which progressively transports each sample to imbalanced clusters under prior and semantic relation constraints, thus generating high-quality and imbalance-aware pseudo-labels.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- Object Affordance Recognition and Grounding via Multi-scale Cross-modal Representation Learning1
A novel approach is proposed that learns an affordance-aware 3D representation and employs a stage-wise inference strategy leveraging the dependency between grounding and classification tasks, enabling inference of full potential affordance areas at a suitable regional scale.
- 1
- 1
- Composing Novel Classes: A Concept-Driven Approach to Generalized Category Discovery1
A novel concept learning framework for GCD is introduced, named ConceptGCD, that categorizes concepts into two types: derivable and underivable from known class concepts, and adopts a stage-wise learning strategy to learn them separately.
- Bayesian Occam’s Razor to Optimize Metamodeling for Complex Biological Systems1
The optimized metamodel delivers more accurate estimates of impaired β-cell dynamics and function in T2D subjects compared to the non-optimized one, underscoring the critical role of optimization in enhancing both model reliability and applicability.
- 1
- 1
- 1
- 3D Object Structure Recovery via Semi-supervised Learning on Videos.1
This work proposes a semi-supervised learning strategy to build a robust 3D object interpreter, which exploits rich object videos for better generalization under large pose variations and noisy 2D key-point estimation.
- Learning to Generate Object Segment Proposals with Multi-modal Cues1
This paper presents a learning-based object segmentation proposal generation method for stereo images that makes use of learned deep features and designed geometric features to represent a region, as well as a learned similarity network to guide the grouping process.
- 1
- 1
- 1
- Caricaturing improves face identity recognition in simulated prosthetic vision1
Caricaturing offers practical benefits for face recognition with the bionic eye by enhancing high-level face shape information and improving both the perceptual individuation and recognition of phoshenised faces.
- 1
- 1
- Hypothesis Testing for Treatment Effect Heterogeneity via an Adaptively Augmented Debiased Estimator–
- –
- –
- Freeze and Cluster: A Simple Baseline for Rehearsal-Free Continual Category Discovery–
This paper proposes a straightforward yet highly effective baseline method that acquires representations using a model specifically trained on the base classes, generates high-quality pseudo-labels through k-means clustering, and trains only the classifier layer.
- A Cell Metamodel Uncovers Mechanistic Drivers of Disease Phenotypes Across Molecular, Cellular, and Tissue Scales–
A graph-based metamodeling framework is developed and applied to construct a comprehensive metamodel of the β-cell, which identifies and characterizes two types of hub cells, each deterministically driving islet activity and synchronization via unique ion channel properties.
- –
- –
- –
- –
- –
- –
- –
- Wavelength division multiplexing reservoir computer using quantum dot lasers with delayed optical feedback–
It is shown that the four-channel WDM RC exhibits a better performance over the single-channel one, with the same number of virtual neurons, and that the cross-gain saturation effect between the multimodes plays a crucial role on the RC performance.
- –
- –
- An Interpretable Multi-scale Deep Network for Structure Localization in Medical Images–
This paper proposes an interpretable multi-scale convolutional networks for structure localization in medical images and demonstrates that the method is capable of generating an interpretable localization process and achieves the state-of-the-art detection performance.
- –
- –
- Studying Object Naming with Online Photos and Caption–
It is found that visual context has interpretable effects on visual naming, by analysing the MSCOCO dataset that has manually annotated objects and captions containing the natural language names for the object.
- –
- –
- Spatial Semantics and Classifier Cascades: The ANU 2011 Multimedia Event Detection System.–
The system uses five audio and visual features, leverages training events with cascaded classifier training, and sees performance improvements with spatial semantic features in the TRECVID-2011 video retrieval benchmark.
- –
- –
- –
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 2026-10-14. 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.