Multi-scale discriminative Region Discovery for Weakly-Supervised Object Localization

TL;DR

This paper introduces a multi-scale discriminative region discovery method for weakly-supervised object localization, leveraging gradient weights across CNN layers to improve localization of multiple and small objects using only image-level labels.

Contribution

It proposes a novel approach that utilizes gradient weights from multiple CNN layers and a parallel sliding window to enhance object localization accuracy in weakly-supervised settings.

Findings

Outperforms previous methods on ILSVRC 2016 with 48.65% Top-1 error

Achieves highest localization accuracy of 0.43 on PASCAL VOC 2012

Provides competitive results on CUB-200-2011 dataset

Abstract

Localizing objects with weak supervision in an image is a key problem of the research in computer vision community. Many existing Weakly-Supervised Object Localization (WSOL) approaches tackle this problem by estimating the most discriminative regions with feature maps (activation maps) obtained by Deep Convolutional Neural Network, that is, only the objects or parts of them with the most discriminative response will be located. However, the activation maps often display different local maximum responses or relatively weak response when one image contains multiple objects with the same type or small objects. In this paper, we propose a simple yet effective multi-scale discriminative region discovery method to localize not only more integral objects but also as many as possible with only image-level class labels. The gradient weights flowing into different convolutional layers of CNN are taken as the input of our method, which is different from previous methods only considering that of the final convolutional layer. To mine more discriminative regions for the task of object localization, the multiple local maximum from the gradient weight maps are leveraged to generate the localization map with a parallel sliding window. Furthermore, multi-scale localization maps from different convolutional layers are fused to produce the final result. We evaluate the proposed method with the foundation of VGGnet on the ILSVRC 2016, CUB-200-2011 and PASCAL VOC 2012 datasets. On ILSVRC 2016, the proposed method yields the Top-1 localization error of 48.65\%, which outperforms previous results by 2.75\%. On PASCAL VOC 2012, our approach achieve the highest localization accuracy of 0.43. Even for CUB-200-2011 dataset, our method still achieves competitive results.