Adaptive Class Weight based Dual Focal Loss for Improved Semantic Segmentation

TL;DR

This paper introduces a Dual Focal Loss (DFL) that improves class imbalance handling in semantic segmentation by focusing on both true and false classes with a learnable scaling parameter, enhancing convergence.

Contribution

The paper presents a novel Dual Focal Loss with a learnable scaling parameter that automatically balances class importance and improves over existing focal loss methods.

Findings

DFL outperforms standard cross entropy in class imbalance scenarios.

The learnable scaling parameter adapts to different datasets.

DFL enhances convergence and segmentation accuracy.

Abstract

In this paper, we propose a Dual Focal Loss (DFL) function, as a replacement for the standard cross entropy (CE) function to achieve a better treatment of the unbalanced classes in a dataset. Our DFL method is an improvement on the recently reported Focal Loss (FL) cross-entropy function, which proposes a scaling method that puts more weight on the examples that are difficult to classify over those that are easy. However, the scaling parameter of FL is empirically set, which is problem-dependent. In addition, like other CE variants, FL only focuses on the loss of true classes. Therefore, no loss feedback is gained from the false classes. Although focusing only on true examples increases probability on true classes and correspondingly reduces probability on false classes due to the nature of the softmax function, it does not achieve the best convergence due to avoidance of the loss on false classes. Our DFL method improves on the simple FL in two ways. Firstly, it takes the idea of FL to focus more on difficult examples than the easy ones, but evaluates loss on both true and negative classes with equal importance. Secondly, the scaling parameter of DFL has been made learnable so that it can tune itself by backpropagation rather than being dependent on manual tuning. In this way, our proposed DFL method offers an auto-tunable loss function that can reduce the class imbalance effect as well as put more focus on both true difficult examples and negative easy examples.