Object Instance Mining for Weakly Supervised Object Detection

TL;DR

This paper proposes an end-to-end object instance mining framework for weakly supervised object detection that detects multiple object instances per image using graph-based information propagation, improving detection performance without extra annotations.

Contribution

Introduces a novel object instance mining framework with graph-based information propagation and reweighted loss for weakly supervised object detection, addressing missing object instances.

Findings

Effective detection of multiple object instances per image.

Improved detection accuracy on VOC datasets.

Demonstrated superiority over existing WSOD methods.

Abstract

Weakly supervised object detection (WSOD) using only image-level annotations has attracted growing attention over the past few years. Existing approaches using multiple instance learning easily fall into local optima, because such mechanism tends to learn from the most discriminative object in an image for each category. Therefore, these methods suffer from missing object instances which degrade the performance of WSOD. To address this problem, this paper introduces an end-to-end object instance mining (OIM) framework for weakly supervised object detection. OIM attempts to detect all possible object instances existing in each image by introducing information propagation on the spatial and appearance graphs, without any additional annotations. During the iterative learning process, the less discriminative object instances from the same class can be gradually detected and utilized for training. In addition, we design an object instance reweighted loss to learn larger portion of each object instance to further improve the performance. The experimental results on two publicly available databases, VOC 2007 and 2012, demonstrate the efficacy of proposed approach.