计算机视觉领域热门研究方向state-of-art算法实时更新

 

细粒度识别

目标检测
Bottom-up Object Detection by Grouping Extreme and Center Points；论文链接；代码链接；
Scale-Aware Trident Networks for Object Detection；论文链接；
Region Proposal by Guided Anchoring；论文链接；
AdaScale: Towards Real-time Video Object Detection Using Adaptive Scaling；论文链接；
Augmentation for small object detection；论文链接；

视觉目标跟踪
SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks；论文链接；
Fast Online Object Tracking and Segmentation: A Unifying Approach；论文链接；
Deeper and Wider Siamese Networks for Real-Time Visual Tracking；论文链接；

多目标跟踪
Multi-Object Tracking with Multiple Cues and Switcher-Aware Classification；论文链接；

语义分割

人体解析
Devil in the Details: Towards Accurate Single and Multiple Human Parsing；论文链接；代码链接

人体位姿估计
Deep High-Resolution Representation Learning for Human Pose Estimation；论文链接；代码链接；项目主页；

密集场景人脸识别（人脸计数）
Selective Refinement Network for High Performance Face Detection；论文链接；
Scale-Aware Attention Network for Crowd Counting；论文链接；
​Improved Selective Refinement Network for Face Detection；论文链接；

超分辨重建
Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network；论文链接；

边缘检测（简笔画/素描）
Photo-Sketching: Inferring Contour Drawings from Images；论文链接；

人脸关键点检测
Super-realtime facial landmark detection and shape fitting by deep regression of shape model parameters；论文链接；代码链接