Interpretation

1. For learning and communication purposes only, if there is any error, please contact me to correct!
2. Do not use cyber violence!
3. Not allowed to spread without permission!
4. QQ group chat: 953606968

Papers

Classic

1. Efficient Graph-Based Image Segmentation
2. Selective Search for Object Recognition
3. RCNN
4. SPPnet
5. Fast RCNN
6. Faster RCNN
7. YOLOv1
8. SSD
9. [YOLO v2]
10. [YOLO v3]
11. Mask RCNN

NMS

1. Learning non-maximum suppression
2. Improving Object Detection With One Line of Code
3. Acquisition of Localization Confidence forAccurate Object Detection
4. Bounding Box Regression with Uncertainty for Accurate Object Detection | 2019

Model Compression and Acceleration

1. 深度神经网络压缩与加速综述
2. Dynamic Deep Neural Networks: Optimizing Accuracy-Efficiency Trade-offs by Selective Execution
3. Dynamic Multi-path Neural Network | 2019
4. Dynamic Neural Network Channel Execution for Efficient Training | 2019

PS: Focus on the acceleration of the training process.

Object Detection with Anchor

Anchor-free Object Detection

1. FCOS: Fully Convolutional One-Stage Object Detection | 2019

Segmentation

1. U-Net
2. 3D CNN
3. 3D U-Net
4. V-Net
5. Panoptic Segmentation | 2019

Structural Optimization

1. Res2Net: A New Multi-scale Backbone Architecture | 2019
2. Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression | 2019

Videos

百度云：