Rethinking Classification and Localization for Object Detection

Abstract

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.

arXiv commentCVPR 2020

References in corpus (10)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• YOLOv3: An Incremental Improvement
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
• DSSD : Deconvolutional Single Shot Detector
• Light-Head R-CNN: In Defense of Two-Stage Object Detector
• CornerNet-Lite: Efficient Keypoint Based Object Detection
• Bottom-up Object Detection by Grouping Extreme and Center Points
• Rethinking Classification and Localization for Cascade R-CNN

Cited by in corpus (5)

• MMDetection: Open MMLab Detection Toolbox and Benchmark
• 1st Place Solutions for OpenImage2019 -- Object Detection and Instance Segmentation
• Revisiting the Sibling Head in Object Detector
• hSDB-instrument: Instrument Localization Database for Laparoscopic and Robotic Surgeries
• GLSD: The Global Large-Scale Ship Database and Baseline Evaluations