Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation
Preprint
Abstract
Automatic segmentation methods are an important advancement in medical image analysis. Machine learning techniques, and deep neural networks in particular, are the state-of-the-art for most medical image segmentation tasks. Issues with class imbalance pose a significant challenge in medical datasets, with lesions often occupying a considerably smaller volume relative to the background. Loss functions used in the training of deep learning algorithms differ in their robustness to class imbalance, with direct consequences for model convergence. The most commonly used loss functions for segmentation are based on either the cross entropy loss, Dice loss or a combination of the two. We propose the Unified Focal loss, a new hierarchical framework that generalises Dice and cross entropy-based losses for handling class imbalance. We evaluate our proposed loss function on five publicly available, class imbalanced medical imaging datasets: CVC-ClinicDB, Digital Retinal Images for Vessel Extraction (DRIVE), Breast Ultrasound 2017 (BUS2017), Brain Tumour Segmentation 2020 (BraTS20) and Kidney Tumour Segmentation 2019 (KiTS19). We compare our loss function performance against six Dice or cross entropy-based loss functions, across 2D binary, 3D binary and 3D multiclass segmentation tasks, demonstrating that our proposed loss function is robust to class imbalance and consistently outperforms the other loss functions. Source code is available at: https://github.com/mlyg/unified-focal-lossKeywords
All Related Versions
This publication has 61 references indexed in Scilit:
- Ensemble U‐net‐based method for fully automated detection and segmentation of renal masses on computed tomography imagesMedical Physics, 2020
- Overfitting of Neural Nets Under Class Imbalance: Analysis and Improvements for Segmentation2019
- Robust Multimodal Brain Tumor Segmentation via Feature Disentanglement and Gated Fusion2019
- AnatomyNet: Deep learning for fast and fully automated whole‐volume segmentation of head and neck anatomyMedical Physics, 2018
- NeXt for neuro‐radiosurgery: A fully automatic approach for necrosis extraction in brain tumor MRI using an unsupervised machine learning techniqueInternational Journal of Imaging Systems and Technology, 2017
- Tversky Loss Function for Image Segmentation Using 3D Fully Convolutional Deep Networks2017
- 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation2016
- U-Net: Convolutional Networks for Biomedical Image Segmentation2015
- DeepOrgan: Multi-level Deep Convolutional Networks for Automated Pancreas Segmentation2015
- Atlas-Based Under-Segmentation2014