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
- Suggested from co-authorship
- Suggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works83 from public data
- Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling1,232
Point-BERT, a new paradigm for learning Transformers to generalize the concept of BERT to 3D point cloud, is presented and it is shown that a pure Transformer architecture attains 93.8% accuracy on ModelNet40 and 83.1% accuracy in the hardest setting of ScanObjectNN, surpassing carefully designed point cloud models with much fewer hand-made designs.
- DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting886
This work presents a new framework for dense prediction by implicitly and explicitly leveraging the pre-trained knowledge from CLIP and converts the original image-text matching problem in CLIP to a pixel- text matching problem and uses the pixel-text score maps to guide the learning of dense prediction models.
- PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers679
A new method is presented that reformulates point cloud completion as a set-to-set translation problem and design a new model, called PoinTr, that adopts a transformer encoder-decoder architecture for point clouds completion that outperforms state-of-the-art methods by a large margin.
- Generative Multimodal Models are In-Context Learners545
This work demonstrates that the task-agnostic in-context learning capabilities of large multimodal models can be significantly enhanced by effective scaling-up, and introduces Emu2, a generative multimodal model with 37 billion parameters, trained on large-scale multimodal sequences with a unified autoregressive objective.
- Runtime Neural Pruning536
A Runtime Neural Pruning (RNP) framework which prunes the deep neural network dynamically at the runtime and preserves the full ability of the original network and conducts pruning according to the input image and current feature maps adaptively.
- COIN: A Large-Scale Dataset for Comprehensive Instructional Video Analysis477
A simple yet effective method to capture the dependencies among different steps, which can be easily plugged into conventional proposal-based action detection methods for localizing important steps in instruction videos is proposed.
- Stochastic Trajectory Prediction via Motion Indeterminacy Diffusion438
This paper presents a new framework to formulate the trajectory prediction task as a reverse process of motion indeterminacy diffusion (MID), in which the history behavior information and the social interactions are encoded as a state embedding and a Transformer-based diffusion model is devised to capture the temporal dependencies of trajectories.
- Structure-Preserving Super Resolution With Gradient Guidance415
A structure-preserving super resolution method which exploits gradient maps of images to guide the recovery in two aspects and proposes a gradient loss which imposes a second-order restriction on the super-resolved images.
- Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification380
This paper analyzes the effect of the learned visual attention on network prediction through counterfactual intervention and maximize the effect to encourage the network to learn more useful attention for fine-grained image recognition.
- Unleashing Text-to-Image Diffusion Models for Visual Perception366
It is shown that vision-language pre-trained diffusion models can be faster adapted to downstream visual perception tasks using the proposed VPD, a new framework that exploits the semantic information of a pre-trained text-to-image diffusion model in visual perception tasks.
- 306
- NerfingMVS: Guided Optimization of Neural Radiance Fields for Indoor Multi-view Stereo293
A new multi-view depth estimation method that utilizes both conventional SfM reconstruction and learning-based priors over the recently proposed neural radiance fields (NeRF), with surprising findings presented on the effectiveness of correspondence-based opti-mization and NeRF-based optimization over the adapted depth priors.
- 217
- FineDiving: A Fine-grained Dataset for Procedure-aware Action Quality Assessment203
This work proposes a procedure-aware approach for action quality assessment, learned by a new Temporal Segmentation Attention module, and proposes to parse pairwise query and exemplar action instances into consecutive steps with diverse semantic and temporal correspondences.
- Deep Face Super-Resolution With Iterative Collaboration Between Attentive Recovery and Landmark Estimation201
Results show the proposed FSR method with iterative collaboration between two recurrent networks which focus on facial image recovery and landmark estimation significantly outperforms state-of-the-art FSR methods in recovering high-quality face images.
- AdaPoinTr: Diverse Point Cloud Completion With Adaptive Geometry-Aware Transformers190
A Transformer encoder-decoder architecture, called PoinTr, is proposed, which reformulates point cloud completion as a set-to-set translation problem and employs a geometry-aware block to model local geometric relationships explicitly and can achieve promising performance with higher throughputs and fewer FLOPs compared with the previous best methods in practice.
- 182
- Group-aware Contrastive Regression for Action Quality Assessment176
It is shown that the relations among videos can provide important clues for more accurate action quality assessment during both training and inference, and a new Contrastive Regression framework is proposed to learn the relative scores by pair-wise comparison, which highlights the differences between videos and guides the models to learning the key hints for assessment.
- 173
- 157
- Insight-V: Exploring Long-Chain Visual Reasoning with Multimodal Large Language Models153
InsightV, an early effort to scalably produce long and robust reasoning data for complex multi-modal tasks, and an effective training pipeline to enhance the reasoning capabilities of multi-modal large language models (MLLMs), is presented.
- Global-Local Bidirectional Reasoning for Unsupervised Representation Learning of 3D Point Clouds152
This work hypothesizes that a powerful representation of a 3D object should model the attributes that are shared between parts and the whole object, and distinguishable from other objects, and proposes to learn point cloud representation by bidirectional reasoning between the local structures at different abstraction hierarchies and the global shape without human supervision.
- PV-RAFT: Point-Voxel Correlation Fields for Scene Flow Estimation of Point Clouds127
Experimental results show that PV-RAFT outperforms state-of-the-art methods by remarkable margins and is evaluated on the FlyingThings3D and KITTI Scene Flow 2015 datasets.
- Structure-Preserving Image Super-Resolution123
A structure-preserving super-resolution (SPSR) method to alleviate the above issue while maintaining the merits of GAN-based methods to generate perceptual-pleasant details, and proposes two self-supervised structure learning methods, contrastive prediction and solving jigsaw puzzles, to train the NSEs.
- SurroundDepth: Entangling Surrounding Views for Self-Supervised Multi-Camera Depth Estimation112
This paper proposes a SurroundDepth method to incorporate the information from multiple surrounding views to predict depth maps across cameras, and adopts the two-frame structure-from-motion to extract scale-aware pseudo depths to pretrain the models.
- 111
- PLOT: Prompt Learning with Optimal Transport for Vision-Language Models86
This work proposes to apply optimal transport to match the vision and text modalities and first model images and the categories with visual and textual feature sets and applies a two-stage optimization strategy to learn the prompts.
- Temporal Coherence or Temporal Motion: Which Is More Critical for Video-Based Person Re-identification?83
This paper proposes a simple yet effective Adversarial Feature Augmentation (AFA) method, which highlights the temporal coherence features by introducing adversarial augmented temporal motion noise.
- Learning Discriminative Aggregation Network for Video-Based Face Recognition and Person Re-identification79
Experimental results show that the proposed discriminative aggregation network method can generate discriminate images from video clips and improve the overall recognition performance in both the speed and the accuracy for video-based face recognition and person re-identification.
- 71
- Chain-of-Spot: Interactive Reasoning Improves Large Vision-Language Models67
This work introduces the Chain-of-Spot (CoS) method, which is described as Interactive Reasoning, a novel approach that enhances feature extraction by focusing on key regions of interest (ROI) within the image, corresponding to the posed questions or instructions, thereby offering multi-granularity image features.
- 66
- Dynamic Spatial Sparsification for Efficient Vision Transformers and Convolutional Neural Networks64
A dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input to accelerate vision Transformers and finds that the idea of adaptive and asymmetric computation can be a general solution for accelerating various architectures.
- RandomRooms: Unsupervised Pre-training from Synthetic Shapes and Randomized Layouts for 3D Object Detection60
A new method called RandomRooms is proposed to generate random layouts of a scene by making use of the objects in the synthetic CAD dataset and learn the 3D scene representation by applying object-level contrastive learning on two random scenes generated from the same set of synthetic objects.
- 59
- Towards Interpretable Deep Metric Learning with Structural Matching58
This paper proposes a structural matching strategy that explicitly aligns the spatial embeddings by computing an optimal matching flow between feature maps of the two images and enables deep models to learn metrics in a more human-friendly way.
- 57
- 52
- 52
- MetaDistiller: Network Self-Boosting via Meta-Learned Top-Down Distillation44
This work proposes that better soft targets with higher compatibil-ity can be generated by using a label generator to fuse the feature maps from deeper stages in a top-down manner, and can employ the meta-learning technique to optimize this label generator.
- 43
- Take-A-Photo: 3D-to-2D Generative Pre-training of Point Cloud Models42
A novel 3D-to-2D generative pre-training method that is adaptable to any point cloud model and effective in boosting the performance of architecture-oriented approaches, achieving state-of-the-art performance when fine-tuning on ScanObjectNN classification and ShapeNet-Part segmentation tasks.
- FLAG3D: A 3D Fitness Activity Dataset with Language Instruction37
FLAG3D, a large-scale 3D fitness activity dataset with language instruction containing 180K sequences of 60 categories, contributes great research value for various challenges, such as cross-domain human action recognition, dynamic human mesh recovery, and language-guided human action generation.
- Back to Reality: Weakly-supervised 3D Object Detection with Shape-guided Label Enhancement34
A weakly-supervised approach for 3D object detection, which makes it possible to train a strong 3D detector with position-level annotations and makes use of synthetic 3D shapes to convert the weak labels into fully-annotated virtual scenes as stronger supervision, and utilizes the perfect virtual labels to complement and refine the real labels.
- TCOVIS: Temporally Consistent Online Video Instance Segmentation29
This paper proposes a novel online method for video instance segmentation, called TCOVIS, which fully exploits the temporal information in a video clip, and achieves state-of-the-art performance on all benchmarks without bells-and-whistles.
- X-3D: Explicit 3D Structure Modeling for Point Cloud Recognition26
X-3D functions by capturing the explicit local structural information within the input 3D space and employing it to produce dynamic kernels with shared weights for all neighborhood points within the current local region, thereby improving the extraction of local features.
- 25
- SparseMM: Head Sparsity Emerges from Visual Concept Responses in MLLMs24
SparseMM is introduced, a KV-Cache optimization strategy that allocates asymmetric computation budgets to heads in LLMs based on their visual scores, leveraging the sparity of visual heads for accelerating the inference of MLLMs.
- 24
- Coarse Correspondences Boost Spatial-Temporal Reasoning in Multimodal Language Model20
This work introduces Coarse Correspondences, a simple lightweight method that enhances MLLMs’ spatial-temporal reasoning with 2D images as input, without modifying the architecture or requiring task-specific fine-tuning.
- 20
- SemAffiNet: Semantic-Affine Transformation for Point Cloud Segmentation19
This work proposes SemAffiNet for point cloud semantic segmentation, which utilizes the attention mechanism in the Transformer module to implicitly and explicitly capture global structural knowledge within local parts for overall comprehension of each category.
- 18
- 18
- Procedure-Aware Action Quality Assessment: Datasets and Performance Evaluation17
A new approach is proposed (i.e., spatial-temporal segmentation attention, STSA) that introduces procedure segmentation to parse an action into consecutive steps, learns powerful representations from these steps by constructing spatial motion attention and procedure-aware cross-attention, and designs a fine-grained contrastive regression to achieve an interpretable scoring mechanism.
- A Roadmap for Big Model17
This paper discusses not only the BM technologies themselves but also the prerequisites for BM training and applications with BMs, dividing the BM review into four parts: Resource, Models, Key Technologies and Application.
- 16
- 16
- AMixer: Adaptive Weight Mixing for Self-attention Free Vision Transformers11
This paper investigates the key differences between vision Transformers and recent all-MLP models and proposes a computationally and parametrically efficient operation named adaptive weight mixing to generate attention weights without token-token interactions and develops a new architecture named AMixer to capture both long-term and short-term spatial dependencies without self-attention.
- Multi-Proxy Wasserstein Classifier for Image Classification11
This work adopts optimal transport theory to calculate a non-uniform matching flow between the elements in the feature map of a sample and the proxies of a class in a closed way, and enables the models to achieve partial matching.
- 10
- 9
- 9
- 3
- Vision Generalist Model: A Survey3
This paper provides a comprehensive overview of the vision generalist models, delving into their characteristics and capabilities within the field, and digs into the design of frameworks that have been proposed in existing research while also introducing the techniques employed to enhance their performance.
- 3
- 2
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
Publication data from OpenAlex; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-16. 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.