Mixed-order self-paced curriculum learning for universal lesion detection

TL;DR

This paper introduces a mixed-order self-paced curriculum learning method that combines uncertainty and loss to improve universal lesion detection in medical images, addressing data imbalance and difficulty estimation issues.

Contribution

The paper proposes a novel mixed-order SCL approach that integrates uncertainty and loss for better difficulty estimation and sample utilization in medical image analysis.

Findings

Improves lesion detection accuracy without extra network complexity.

Effectively handles class imbalance and data scarcity in medical datasets.

Enhances existing ULD methods with a simple, effective curriculum learning strategy.

Abstract

Self-paced curriculum learning (SCL) has demonstrated its great potential in computer vision, natural language processing, etc. During training, it implements easy-to-hard sampling based on online estimation of data difficulty. Most SCL methods commonly adopt a loss-based strategy of estimating data difficulty and deweighting the `hard' samples in the early training stage. While achieving success in a variety of applications, SCL stills confront two challenges in a medical image analysis task, such as universal lesion detection, featuring insufficient and highly class-imbalanced data: (i) the loss-based difficulty measurer is inaccurate; ii) the hard samples are under-utilized from a deweighting mechanism. To overcome these challenges, in this paper we propose a novel mixed-order self-paced curriculum learning (Mo-SCL) method. We integrate both uncertainty and loss to better estimate difficulty online and mix both hard and easy samples in the same mini-batch to appropriately alleviate the problem of under-utilization of hard samples. We provide a theoretical investigation of our method in the context of stochastic gradient descent optimization and extensive experiments based on the DeepLesion benchmark dataset for universal lesion detection (ULD). When applied to two state-of-the-art ULD methods, the proposed mixed-order SCL method can provide a free boost to lesion detection accuracy without extra special network designs.