Pattern Recognition and Computer Vision 4th Chinese Conference, PRCV 2021, Beijing, China, October 29 – November 1, 2021, Proceedings, Part IV

The 4-volume set LNCS 13019, 13020, 13021 and 13022 constitutes the refereed proceedings of the 4th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2021, held in Beijing, China, in October-November 2021.

The 201 full papers presented were carefully reviewed and selected from 513 submissions. The papers have been organized in the following topical sections: Object Detection, Tracking and Recognition; Computer Vision, Theories and Applications, Multimedia Processing and Analysis; Low-level Vision and Image Processing; Biomedical Image Processing and Analysis; Machine Learning, Neural Network and Deep Learning, and New Advances in Visual Perception and Understanding.

Machine Learning, Neural Network and Deep Learning.- Edge-Wise One-Level Global Pruning on NAS generated networks.- Convolution Tells Where to Look.- Robust Single-step Adversarial Training with Regularizer.- Texture-guided U-Net for OCT-to-OCTA Generation.- Learning Key Actors and Their Interactions for Group Activity Recognition.- Attributed Non-negative Matrix Multi-Factorization for Data Representation.- Improved Categorical Cross-Entropy Loss for Training Deep Neural Networks with Noisy Labels.- A Residual Correction Approach for Semi-supervised Semantic Segmentation.- Hypergraph Convolutional Network with Hybrid Higher-order Neighbors.- Text-Aware Single Image Specular Highlight Removal.- Minimizing Wasserstein-1 Distance by Quantile Regression for GANs Model.- A Competition of Shape and Texture Bias by Multi-View Image Representation.- Learning Indistinguishable and Transferable Adversarial Examples.- Efficient Object Detection and Classification of Ground Objects fromThermal Infrared Remote Sensing Image Based on Deep Learning.- MEMA-NAS: Memory-Efficient Multi-Agent Neural Architecture Search.- Adversarial Decoupling for Weakly Supervised Semantic Segmentation.- Towards End-to-End Embroidery Style Generation: A Paired Dataset and Benchmark.- Efficient and real-time particle detection via encoder-decoder network.- Flexible Projection Search using Optimal Re-weighted Adjacency for Unsupervised Manifold Learning .- Fabric Defect Detection via Multi-scale Feature Fusion-based Saliency.- Improving Adversarial Robustness of Detectors via Objectness Regularization.- IPE Transformer for Depth Completion with Input-Aware Positional Embeddings.- Enhanced Multi-view Matrix Factorization with Shared Representation.- Multi-level Residual Attention Network for Speckle Suppression.- Suppressing Style-Sensitive Features via Randomly Erasing for Domain Generalizable Semantic Segmentation.- MAGAN: Multi-Attention Generative Adversarial Networks for Text-to-Image Generation.- Dual Attention Based Network with Hierarchical ConvLSTM for Video Object Segmentation.- Distance-based Class Activation Map for Metric Learning.- Reading Pointer Meter through One Stage End-to-End Deep Regression.- Deep Architecture Compression with Automatic Clustering of Similar Neurons.- Attention Guided Spatio-temporal Artifacts Extraction  for Deepfake Detection.- Learn the Approximation Distribution of Sparse Coding with Mixture Sparsity Network.- Anti-occluded person re-identification via pose restoration and dual channel feature distance measurement.- Dynamic Runtime Feature Map Pruning.- Special Session: New Advances in Visual Perception and Understanding.- Multi-Branch Graph Network for Learning Human-Object Interaction.- FDEA: Face Dataset with Ethnicity Attribute.- TMD-FS: Improving Few-Shot Object Detection with Transformer Multi-modal Directing.- Feature Matching Network for Weakly-Supervised Temporal Action Localization.- LiDAR-based symmetrical guidance for 3D Object Detection.- Few-shot Segmentation via Complementary Prototype Learning and Cascaded Refinement.- Couple Double-Stage FPNs with Single Pipe-line for solar speckle images deblurring.- Multi-scale Image Partitioning and Saliency Detection for Single Image Blind Deblurring.- CETransformer: Casual Effect Estimation via Transformer Based Representation Learning.- An Efficient Polyp Detection Framework with Suspicious Targets Assisted Training.- Invertible Image Compressive Sensing.- Gradient-free Neural Network Training Based on Deep Dictionary Learning with the Log Regularizer.