KL-Divergence-Based Region Proposal Network for Object Detection

TL;DR

This paper introduces a novel region proposal network that uses KL-divergence to unify classification and bounding box regression, improving object detection accuracy.

Contribution

It redefines the RPN as a KL-divergence minimization problem, integrating uncertainty in bounding box predictions for better proposals.

Findings

Achieves 2.6% AP improvement on MS COCO with Faster R-CNN.

Achieves 2.0% AP improvement on MS COCO with R-FCN.

Demonstrates effectiveness of KL-divergence in region proposal learning.

Abstract

The learning of the region proposal in object detection using the deep neural networks (DNN) is divided into two tasks: binary classification and bounding box regression task. However, traditional RPN (Region Proposal Network) defines these two tasks as different problems, and they are trained independently. In this paper, we propose a new region proposal learning method that considers the bounding box offset's uncertainty in the objectness score. Our method redefines RPN to a problem of minimizing the KL-divergence, difference between the two probability distributions. We applied KL-RPN, which performs region proposal using KL-Divergence, to the existing two-stage object detection framework and showed that it can improve the performance of the existing method. Experiments show that it achieves 2.6% and 2.0% AP improvements on MS COCO test-dev in Faster R-CNN with VGG-16 and R-FCN with ResNet-101 backbone, respectively.