Adaptive Feature Pooling

Introduced by Liu et al. in Path Aggregation Network for Instance Segmentation

Adaptive Feature Pooling pools features from all levels for each proposal in object detection and fuses them for the following prediction. For each proposal, we map them to different feature levels. Following the idea of Mask R-CNN, RoIAlign is used to pool feature grids from each level. Then a fusion operation (element-wise max or sum) is utilized to fuse feature grids from different levels.

The motivation for this technique is that in an FPN we assign proposals to different feature levels based on the size of proposals, which could be suboptimal if images with small differences are assigned to different levels, or if the importance of features is not strongly correlated to their level which they belong.

Source: Path Aggregation Network for Instance Segmentation

Latest Papers

Pyramid Attention Networks for Image Restoration
| Yiqun MeiYuchen FanYulun ZhangJiahui YuYuqian ZhouDing LiuYun FuThomas S. HuangHumphrey Shi
YOLOv4: Optimal Speed and Accuracy of Object Detection
| Alexey BochkovskiyChien-Yao WangHong-Yuan Mark Liao
CSPNet: A New Backbone that can Enhance Learning Capability of CNN
| Chien-Yao WangHong-Yuan Mark LiaoI-Hau YehYueh-Hua WuPing-Yang ChenJun-Wei Hsieh
PANet: Few-Shot Image Semantic Segmentation with Prototype Alignment
| Kaixin WangJun Hao LiewYingtian ZouDaquan ZhouJiashi Feng
iSAID: A Large-scale Dataset for Instance Segmentation in Aerial Images
Syed Waqas ZamirAditya AroraAkshita GuptaSalman KhanGuolei SunFahad Shahbaz KhanFan ZhuLing ShaoGui-Song XiaXiang Bai
Path Aggregation Network for Instance Segmentation
| Shu LiuLu QiHaifang QinJianping ShiJiaya Jia


Object Detection 4 21.05%
Semantic Segmentation 3 15.79%
Real-Time Object Detection 2 10.53%
Instance Segmentation 2 10.53%
Demosaicking 1 5.26%
Denoising 1 5.26%
Image Denoising 1 5.26%
Image Restoration 1 5.26%
Super-Resolution 1 5.26%