A patch-based architecture for multi-label classification from single label annotations

TL;DR

This paper introduces a patch-based attention architecture for multi-label image classification using only single label annotations, enabling training from scratch and improving negative example estimation.

Contribution

The paper presents a novel patch-based architecture with attention and self-similarity strategies for positive and unlabeled learning in multi-label classification.

Findings

Can be trained from scratch without pre-training

Effective estimation of negative examples using patch self-similarities

Outperforms related methods requiring pre-training

Abstract

In this paper, we propose a patch-based architecture for multi-label classification problems where only a single positive label is observed in images of the dataset. Our contributions are twofold. First, we introduce a light patch architecture based on the attention mechanism. Next, leveraging on patch embedding self-similarities, we provide a novel strategy for estimating negative examples and deal with positive and unlabeled learning problems. Experiments demonstrate that our architecture can be trained from scratch, whereas pre-training on similar databases is required for related methods from the literature.