MARIO: A Mixed Annotation Framework For Polyp Segmentation

TL;DR

MARIO is a versatile polyp segmentation framework that effectively utilizes multiple types of annotations, including weak labels, to improve performance and reduce reliance on costly, fully annotated datasets.

Contribution

MARIO introduces a mixed supervision approach that leverages diverse annotation types with tailored loss functions, expanding usable data and enhancing segmentation accuracy.

Findings

Outperforms existing methods on five benchmark datasets.

Effectively utilizes weak and cheap annotations.

Balances trade-offs between different supervision forms.

Abstract

Existing polyp segmentation models are limited by high labeling costs and the small size of datasets. Additionally, vast polyp datasets remain underutilized because these models typically rely on a single type of annotation. To address this dilemma, we introduce MARIO, a mixed supervision model designed to accommodate various annotation types, significantly expanding the range of usable data. MARIO learns from underutilized datasets by incorporating five forms of supervision: pixel-level, box-level, polygon-level, scribblelevel, and point-level. Each form of supervision is associated with a tailored loss that effectively leverages the supervision labels while minimizing the noise. This allows MARIO to move beyond the constraints of relying on a single annotation type. Furthermore, MARIO primarily utilizes dataset with weak and cheap annotations, reducing the dependence on large-scale, fully annotated ones. Experimental results across five benchmark datasets demonstrate that MARIO consistently outperforms existing methods, highlighting its efficacy in balancing trade-offs between different forms of supervision and maximizing polyp segmentation performance