整理：AI算法与图像处理

CVPR2022论文和代码整理：https://github.com/DWCTOD/CVPR2022-Papers-with-Code-Demo

欢迎关注：

StyleT2I: Toward Compositional and High-Fidelity Text-to-Image Synthesis

• 论文/Paper: http://arxiv.org/pdf/2203.15799

• 代码/Code: https://github.com/zhihengli-UR/StyleT2I

CHEX: CHannel EXploration for CNN Model Compression

• 论文/Paper: http://arxiv.org/pdf/2203.15794

• 代码/Code: None

FisherMatch: Semi-Supervised Rotation Regression via Entropy-based Filtering

• 论文/Paper: http://arxiv.org/pdf/2203.15765

• 代码/Code: None

Integrative Few-Shot Learning for Classification and Segmentation

• 论文/Paper: http://arxiv.org/pdf/2203.15712

• 代码/Code: None

OakInk: A Large-scale Knowledge Repository for Understanding Hand-Object Interaction

• 论文/Paper: http://arxiv.org/pdf/2203.15709

• 代码/Code: None

EnvEdit: Environment Editing for Vision-and-Language Navigation

• 论文/Paper: http://arxiv.org/pdf/2203.15685

• 代码/Code: None

Exploring Frequency Adversarial Attacks for Face Forgery Detection

• 论文/Paper: http://arxiv.org/pdf/2203.15674

• 代码/Code: None

PoseTriplet: Co-evolving 3D Human Pose Estimation, Imitation, and Hallucination under Self-supervision

• 论文/Paper: http://arxiv.org/pdf/2203.15625

• 代码/Code: None

Learning a Structured Latent Space for Unsupervised Point Cloud Completion

• 论文/Paper: http://arxiv.org/pdf/2203.15580

• 代码/Code: None

BARC: Learning to Regress 3D Dog Shape from Images by Exploiting Breed Information

• 论文/Paper: http://arxiv.org/pdf/2203.15536

• 代码/Code: None

OSOP: A Multi-Stage One Shot Object Pose Estimation Framework

• 论文/Paper: http://arxiv.org/pdf/2203.15533

• 代码/Code: None

Learning Structured Gaussians to Approximate Deep Ensembles

• 论文/Paper: http://arxiv.org/pdf/2203.15485

• 代码/Code: None

Long-term Video Frame Interpolation via Feature Propagation

• 论文/Paper: http://arxiv.org/pdf/2203.15427

• 代码/Code: None

Quantifying Societal Bias Amplification in Image Captioning

• 论文/Paper: http://arxiv.org/pdf/2203.15395

• 代码/Code: None

Alignment-Uniformity aware Representation Learning for Zero-shot Video Classification

• 论文/Paper: http://arxiv.org/pdf/2203.15381

• 代码/Code: https://github.com/ShipuLoveMili/CVPR2022-AURL

A Style-aware Discriminator for Controllable Image Translation

• 论文/Paper: http://arxiv.org/pdf/2203.15375

• 代码/Code: None

Self-Supervised Image Representation Learning with Geometric Set Consistency

• 论文/Paper: http://arxiv.org/pdf/2203.15361

• 代码/Code: None

Nested Collaborative Learning for Long-Tailed Visual Recognition

• 论文/Paper: http://arxiv.org/pdf/2203.15359

• 代码/Code: None

Online Continual Learning on a Contaminated Data Stream with Blurry Task Boundaries

• 论文/Paper: http://arxiv.org/pdf/2203.15355

• 代码/Code: None

SIOD: Single Instance Annotated Per Category Per Image for Object Detection

• 论文/Paper: http://arxiv.org/pdf/2203.15353

• 代码/Code: None

Task-specific Inconsistency Alignment for Domain Adaptive Object Detection

• 论文/Paper: http://arxiv.org/pdf/2203.15345

• 代码/Code: None

Balanced Multimodal Learning via On-the-fly Gradient Modulation

• 论文/Paper: http://arxiv.org/pdf/2203.15332

• 代码/Code: None

CNN Filter DB: An Empirical Investigation of Trained Convolutional Filters

• 论文/Paper: http://arxiv.org/pdf/2203.15331

• 代码/Code: None

Dressing in the Wild by Watching Dance Videos

• 论文/Paper: http://arxiv.org/pdf/2203.15320

• 代码/Code: None

Eigenlanes: Data-Driven Lane Descriptors for Structurally Diverse Lanes

• 论文/Paper: http://arxiv.org/pdf/2203.15302

• 代码/Code: None

Uncertainty-Aware Adaptation for Self-Supervised 3D Human Pose Estimation

• 论文/Paper: http://arxiv.org/pdf/2203.15293

• 代码/Code: None

MAT: Mask-Aware Transformer for Large Hole Image Inpainting

• 论文/Paper: http://arxiv.org/pdf/2203.15270

• 代码/Code: None

Eigencontours: Novel Contour Descriptors Based on Low-Rank Approximation

• 论文/Paper: http://arxiv.org/pdf/2203.15259

• 代码/Code: None

Pop-Out Motion: 3D-Aware Image Deformation via Learning the Shape Laplacian

• 论文/Paper: http://arxiv.org/pdf/2203.15235

• 代码/Code: None

Zero-Query Transfer Attacks on Context-Aware Object Detectors

• 论文/Paper: http://arxiv.org/pdf/2203.15230

• 代码/Code: None

Temporal Feature Alignment and Mutual Information Maximization for Video-Based Human Pose Estimation

• 论文/Paper: http://arxiv.org/pdf/2203.15227

• 代码/Code: None

Affine Medical Image Registration with Coarse-to-Fine Vision Transformer

• 论文/Paper: http://arxiv.org/pdf/2203.15216

• 代码/Code: https://github.com/cwmok/C2FViT

Camera-Conditioned Stable Feature Generation for Isolated Camera Supervised Person Re-IDentification

• 论文/Paper: http://arxiv.org/pdf/2203.15210

• 代码/Code: None

SPAct: Self-supervised Privacy Preservation for Action Recognition

• 论文/Paper: http://arxiv.org/pdf/2203.15205

• 代码/Code: None

SimT: Handling Open-set Noise for Domain Adaptive Semantic Segmentation

• 论文/Paper: http://arxiv.org/pdf/2203.15202

• 代码/Code: None

3D Shape Reconstruction from 2D Images with Disentangled Attribute Flow

• 论文/Paper: http://arxiv.org/pdf/2203.15190

• 代码/Code: None

ASM-Loc: Action-aware Segment Modeling for Weakly-Supervised Temporal Action Localization

• 论文/Paper: http://arxiv.org/pdf/2203.15187

• 代码/Code: None

Long-term Visual Map Sparsification with Heterogeneous GNN

• 论文/Paper: http://arxiv.org/pdf/2203.15182

• 代码/Code: None

Unified Transformer Tracker for Object Tracking

• 论文/Paper: http://arxiv.org/pdf/2203.15175

• 代码/Code: None

Towards End-to-End Unified Scene Text Detection and Layout Analysis

• 论文/Paper: http://arxiv.org/pdf/2203.15143

• 代码/Code: None

Text2Pos: Text-to-Point-Cloud Cross-Modal Localization

• 论文/Paper: http://arxiv.org/pdf/2203.15125

• 代码/Code: None

LiDAR Snowfall Simulation for Robust 3D Object Detection

• 论文/Paper: http://arxiv.org/pdf/2203.15118

• 代码/Code: None

Rethinking Semantic Segmentation: A Prototype View

• 论文/Paper: http://arxiv.org/pdf/2203.15102

• 代码/Code: None

X-Pool: Cross-Modal Language-Video Attention for Text-Video Retrieval

• 论文/Paper: http://arxiv.org/pdf/2203.15086

• 代码/Code: https://github.com/layer6ai-labs/xpool

Few-Shot Object Detection with Fully Cross-Transformer

• 论文/Paper: http://arxiv.org/pdf/2203.15021

• 代码/Code: None

Registering Explicit to Implicit: Towards High-Fidelity Garment mesh Reconstruction from Single Images

• 论文/Paper: http://arxiv.org/pdf/2203.15007

• 代码/Code: None

Clean Implicit 3D Structure from Noisy 2D STEM Images

• 论文/Paper: http://arxiv.org/pdf/2203.15434

• 代码/Code: None

Equivariance Allows Handling Multiple Nuisance Variables When Analyzing Pooled Neuroimaging Datasets

• 论文/Paper: http://arxiv.org/pdf/2203.15234

• 代码/Code: None