Fuzzy Overclustering: Semi-Supervised Classification of Fuzzy Labels with Overclustering and Inverse Cross-Entropy

TL;DR

This paper introduces a semi-supervised overclustering framework with a novel loss function to better handle fuzzy labels in classification tasks, especially in underwater image data, improving consistency and substructure detection.

Contribution

It presents a new overclustering-based semi-supervised learning method with a novel loss for fuzzy labels, outperforming existing methods on real-world underwater plankton data.

Findings

Outperforms previous semi-supervised methods on fuzzy label data

Achieves 5-10% more consistent predictions of substructures

Effectively detects substructures within ambiguous labels

Abstract

Deep learning has been successfully applied to many classification problems including underwater challenges. However, a long-standing issue with deep learning is the need for large and consistently labeled datasets. Although current approaches in semi-supervised learning can decrease the required amount of annotated data by a factor of 10 or even more, this line of research still uses distinct classes. For underwater classification, and uncurated real-world datasets in general, clean class boundaries can often not be given due to a limited information content in the images and transitional stages of the depicted objects. This leads to different experts having different opinions and thus producing fuzzy labels which could also be considered ambiguous or divergent. We propose a novel framework for handling semi-supervised classifications of such fuzzy labels. It is based on the idea of overclustering to detect substructures in these fuzzy labels. We propose a novel loss to improve the overclustering capability of our framework and show the benefit of overclustering for fuzzy labels. We show that our framework is superior to previous state-of-the-art semi-supervised methods when applied to real-world plankton data with fuzzy labels. Moreover, we acquire 5 to 10\% more consistent predictions of substructures.