Research


MemBN: Robust Test-Time Adaptation via Batch Norm with Statistics Memory

PROJECT PAGE

Efficient and Versatile Robust Fine-Tuning of Zero-shot Models

PROJECT PAGE

In Defense of Lazy Visual Grounding for Open-Vocabulary Semantic Segmentation

PROJECT PAGE

Online Temporal Action Localization with Memory-Augmented Transformer

PROJECT PAGE

Towards More Practivcal Group Activity Detection: A New Benchmark and Model

PROJECT PAGE

Classification Matters: Improving Video Action Detection with Class-Specific Attention

PROJECT PAGE

Burst Image Super-Resolution with Base Frame Selection

PROJECT PAGE

Learning SO(3)-Invariant Semantic Correspondence via Local Shape Transform

PROJECT PAGE

Contrastive Mean-Shift Learning for Generalized Category Discovery

PROJECT PAGE

Learning Correlation Structures for Vision Transformers

PROJECT PAGE

MoReVQA: Exploring Modular Reasoning Models for Video Question Answering

PROJECT PAGE

Self-supervised Learning of Semantic Correspondence Using Web Videos

PROJECT PAGE

Efficient Semantic Matching with Hypercolumn Correlation

PROJECT PAGE

NVS-Adapter: Plug-and-Play Novel View Synthesis from a Single Image

PROJECT PAGE

Activity Grammars for Temporal Action Segmentation

PROJECT PAGE

Shatter and Gather: Learning Referring Image Segmentation with Text Supervision

PROJECT PAGE

Leveraging Proxy of Training Data for Test-Time Adaptation

PROJECT PAGE

Scaling up GANs for Text-to-Image Synthesis

PROJECT PAGE

Improving Cross-Modal Retrieval with Set of Diverse Embeddings

PROJECT PAGE

WEDGE: Web-Image Assisted Domain Generalization for Semantic Segmentation

PROJECT PAGE

Combating Label Distribution Shift for Active Domain Adaptation

GITHUB PAGE

Relational Context Learning for Human-Object Interaction Detection

PROJECT PAGE

HIER: Metric Learning Beyond Class Labels via Hierarchical Regularization

PROJECT PAGE

Devil's on the Edges: Selective Quad Attention for Scene Graph Generation

PROJECT PAGE

Learning Rotation-Equivariant Features for Visual Correspondence

PROJECT PAGE

3D Scene Painting via Semantic Image Synthesis

PROJECT PAGE

Style Neophile: Constantly Seeking Novel Styles for Domain Generalization

PROJECT PAGE

Peripheral Vision Transformer

PROJECT PAGE

Future Transformer for Long-Term Action Anticipation

PROJECT PAGE

TransforMatcher: Match-to-Match Attention for Semantic Correspondence

PROJECT PAGE

Self-Taught Metric Learning without Labels

PROJECT PAGE

Integrative Few-Shot Learning for Classification and Segmentation

PROJECT PAGE

Fast Point Transformer

PROJECT PAGE

FIFO: Learning Fog-invariant Features for Foggy Scene Segmentation

PROJECT PAGE

DENse and DIverse symmetry dataset (DENDI)

PROJECT PAGE

Reflection and Rotation Symmetry Detection via Equivariant Learning

PROJECT PAGE

Detector-Free Weakly Supervised Group Activity Recognition

PROJECT PAGE

ReSTR: Convolution-free Referring Image Segmentation Using Transformers

PROJECT PAGE

Self-Supervised Equivariant Learning for Oriented Keypoint Detection

PROJECT PAGE

Semi-supervised Semantic Segmentation with Error Localization Network

PROJECT PAGE

Relational Self-Attention: What's Missing in Attention for Video Understanding

PROJECT PAGE

Deep Hough Voting for Robust Global Registration

PROJECT PAGE

Self-Calibrating Neural Radiance Fields

PROJECT PAGE

Learning to Discover Reflection Symmetry via Polar Matching Convolution

PROJECT PAGE

Learning Self-Similarity in Space and Time as Generalized Motion for Video Action Recognition

PROJECT PAGE

Relational Embedding for Few-Shot Classification

PROJECT PAGE

ASMR: Learning Attribute-Based Person Search with Adaptive Semantic Margin Regularizer

PROJECT PAGE

Hypercorrelation Squeeze for Few-Shot Segmentation

PROJECT PAGE

Embedding Transfer with Label Relaxation for Improved Metric Learning

PROJECT PAGE

Convolutional Hough Matching Networks

PROJECT PAGE

URIE: Universal Image Enhancement for Visual Recognition
in the Wild

PROJECT PAGE

MotionSqueeze:
Neural Motion Feature Learning for Video Understanding

PROJECT PAGE

Learning to Compose Hypercolumns for Visual Correspondence

PROJECT PAGE

Proxy Anchor Loss for Deep Metric Learning

PROJECT PAGE

SPair-71k:
A Large-scale Benchmark for Semantic Correspondence

PROJECT PAGE

Hyperpixel Flow:
Semantic Correspondence with Multi-layer Neural Features

PROJECT PAGE

Deep Metric Learning Beyond Binary Supervision

PROJECT PAGE

Relational Knowledge Distillation

PROJECT PAGE

Attentive Semantic Alignment with Offset-Aware Correlation Kernels

PROJECT PAGE

Visual Reference Resolution using Attention Memory for Visual Dialog

PROJECT PAGE

MarioQA: Answering Questions by Watching Gameplay Videos

PROJECT PAGE

Weakly Supervised Semantic Segmentation using Web-Crawled Videos

PROJECT PAGE

Superpixel-based Tracking-by-Segmentation using Markov Chains

PROJECT PAGE

Text-guided Attention Model for Image Captioning

PROJECT PAGE

Training Recurrent Answering Units with Joint Loss Minimization for VQA

PROJECT PAGE

Superpixel segmentation by constrained minimax label propagation

PROJECT PAGE

TransferNet: Transfer learning for semantic segmentation

PROJECT PAGE

MDNet for visual tracking "VOT2015 Winner"

PROJECT PAGE

DPPnet for image question answering

PROJECT PAGE

Unsupervised Co-activity Detection from Multiple Videos using Absorbing Markov Chain

PROJECT PAGE

DecoupledNet for semi-supervised semantic segmentation

PROJECT PAGE

DeconvNet for semantic segmentation

PROJECT PAGE

Tracking-by-segmentation using online GBDT

PROJECT PAGE

Online tracking by learning discriminative saliency map with CNN

PROJECT PAGE

Beyond chain models for visual tracking: A Trilogy

PROJECT PAGE

Event detection

PROJECT PAGE

Joint human segmentation and pose tracking

PUBLICATION DATASET

Tracking with occlusion reasoning

PROJECT PAGE

Generalized background subtraction

PROJECT PAGE

Fast nearest neighbor search

PROJECT PAGE