Progressive Representation Adaptation for Weakly Supervised Object Localization

TL;DR

This paper introduces a progressive adaptation approach for weakly supervised object localization, improving detection accuracy by transferring learned representations and refining object proposals through classification and detection adaptation steps.

Contribution

It proposes a novel two-step progressive adaptation method that enhances weakly supervised object localization by reducing noise and refining proposals through classification and detection adaptation.

Findings

Outperforms state-of-the-art methods on PASCAL VOC and ILSVRC datasets.

Effectively reduces noise in object proposals.

Improves localization accuracy with refined detection.

Abstract

We address the problem of weakly supervised object localization where only image-level annotations are available for training object detectors. Numerous methods have been proposed to tackle this problem through mining object proposals. However, a substantial amount of noise in object proposals causes ambiguities for learning discriminative object models. Such approaches are sensitive to model initialization and often converge to undesirable local minimum solutions. In this paper, we propose to overcome these drawbacks by progressive representation adaptation with two main steps: 1) classification adaptation and 2) detection adaptation. In classification adaptation, we transfer a pre-trained network to a multi-label classification task for recognizing the presence of a certain object in an image. Through the classification adaptation step, the network learns discriminative representations that are specific to object categories of interest. In detection adaptation, we mine class-specific object proposals by exploiting two scoring strategies based on the adapted classification network. Class-specific proposal mining helps remove substantial noise from the background clutter and potential confusion from similar objects. We further refine these proposals using multiple instance learning and segmentation cues. Using these refined object bounding boxes, we fine-tune all the layer of the classification network and obtain a fully adapted detection network. We present detailed experimental validation on the PASCAL VOC and ILSVRC datasets. Experimental results demonstrate that our progressive representation adaptation algorithm performs favorably against the state-of-the-art methods.