Minimizing Supervision in Multi-label Categorization

TL;DR

This paper proposes an active learning approach for multi-label image classification that significantly reduces supervision needs by selecting informative samples, achieving near full supervision performance with only 10-20% of labeled data.

Contribution

It introduces a novel sample selection criterion that improves active learning efficiency in multi-label categorization, outperforming existing methods across datasets.

Findings

Achieves over 98% of fully supervised accuracy with 20% labeled data.

Outperforms baseline metrics on multiple benchmark datasets.

Provides a robust sample selection strategy for reduced supervision.

Abstract

Multiple categories of objects are present in most images. Treating this as a multi-class classification is not justified. We treat this as a multi-label classification problem. In this paper, we further aim to minimize the supervision required for providing supervision in multi-label classification. Specifically, we investigate an effective class of approaches that associate a weak localization with each category either in terms of the bounding box or segmentation mask. Doing so improves the accuracy of multi-label categorization. The approach we adopt is one of active learning, i.e., incrementally selecting a set of samples that need supervision based on the current model, obtaining supervision for these samples, retraining the model with the additional set of supervised samples and proceeding again to select the next set of samples. A crucial concern is the choice of the set of samples. In doing so, we provide a novel insight, and no specific measure succeeds in obtaining a consistently improved selection criterion. We, therefore, provide a selection criterion that consistently improves the overall baseline criterion by choosing the top k set of samples for a varied set of criteria. Using this criterion, we are able to show that we can retain more than 98% of the fully supervised performance with just 20% of samples (and more than 96% using 10%) of the dataset on PASCAL VOC 2007 and 2012. Also, our proposed approach consistently outperforms all other baseline metrics for all benchmark datasets and model combinations.