Global Weighted Average Pooling Bridges Pixel-level Localization and Image-level Classification

TL;DR

This paper introduces a novel global weighted average pooling (GWAP) method for improved pixel-level localization and image classification using only image-level labels, and demonstrates its effectiveness in multi-task learning for object detection.

Contribution

The paper proposes class-agnostic and class-specific GWAP modules and integrates them into a multi-task framework with R-FCN, enhancing localization and detection with limited supervision.

Findings

GWAP better captures foreground object regions.

The multi-task framework improves object detection accuracy.

The approach performs well with limited bounding box annotations.

Abstract

In this work, we first tackle the problem of simultaneous pixel-level localization and image-level classification with only image-level labels for fully convolutional network training. We investigate the global pooling method which plays a vital role in this task. Classical global max pooling and average pooling methods are hard to indicate the precise regions of objects. Therefore, we revisit the global weighted average pooling (GWAP) method for this task and propose the class-agnostic GWAP module and the class-specific GWAP module in this paper. We evaluate the classification and pixel-level localization ability on the ILSVRC benchmark dataset. Experimental results show that the proposed GWAP module can better capture the regions of the foreground objects. We further explore the knowledge transfer between the image classification task and the region-based object detection task. We propose a multi-task framework that combines our class-specific GWAP module with R-FCN. The framework is trained with few ground truth bounding boxes and large-scale image-level labels. We evaluate this framework on PASCAL VOC dataset. Experimental results show that this framework can use the data with only image-level labels to improve the generalization of the object detection model.