Rethinking Classification and Localization for Object Detection
This work addresses a specific bottleneck in object detection by optimizing head structures, offering incremental improvements for computer vision applications.
The paper tackled the problem of understanding and optimizing head structures for classification and localization in object detection, finding that fully connected heads are better for classification and convolutional heads for localization, and proposed a Double-Head method that improved AP by +3.5 and +2.8 on MS COCO with ResNet-50 and ResNet-101 backbones, respectively.
Two head structures (i.e. fully connected head and convolution head) have been widely used in R-CNN based detectors for classification and localization tasks. However, there is a lack of understanding of how does these two head structures work for these two tasks. To address this issue, we perform a thorough analysis and find an interesting fact that the two head structures have opposite preferences towards the two tasks. Specifically, the fully connected head (fc-head) is more suitable for the classification task, while the convolution head (conv-head) is more suitable for the localization task. Furthermore, we examine the output feature maps of both heads and find that fc-head has more spatial sensitivity than conv-head. Thus, fc-head has more capability to distinguish a complete object from part of an object, but is not robust to regress the whole object. Based upon these findings, we propose a Double-Head method, which has a fully connected head focusing on classification and a convolution head for bounding box regression. Without bells and whistles, our method gains +3.5 and +2.8 AP on MS COCO dataset from Feature Pyramid Network (FPN) baselines with ResNet-50 and ResNet-101 backbones, respectively.