A new family of instance-level loss functions for improving instance-level segmentation and detection of white matter hyperintensities in routine clinical brain MRI

• Published in: Computers in biology and medicine Vol. 174; p. 108414
• Main Authors: Rachmadi, Muhammad Febrian, Byra, Michal, Skibbe, Henrik
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.05.2024; Elsevier Limited
• Subjects: Alzheimer's disease; Biology; Biomedical data; Brain; Brain lesions; Brain research; Computers; Datasets; Ensemble inference; Histopathology; Image processing; Image segmentation; Inference; Instance segmentation; Instance-level detection loss; Instance-level segmentation loss; Magnetic resonance imaging; Medical imaging; Neural networks; Neurodegenerative diseases; Neuroimaging; Object recognition; Pixels; Substantia alba; White matter hyperintensities; Instance-level detection loss; White matter hyperintensities; Brain lesions; Instance-level segmentation loss; Ensemble inference
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2024.108414

In this study, we introduce “instance loss functions”, a new family of loss functions designed to enhance the training of neural networks in the instance-level segmentation and detection of objects in biomedical image data, particularly those of varied numbers and sizes. Intended to be utilized conjointly with traditional loss functions, these proposed functions, prioritize object instances over pixel-by-pixel comparisons. The specific functions, the instance segmentation loss (Linstance), the instance center loss (Lcenter), the false instance rate loss (Lfalse), and the instance proximity loss (Lproximity), serve distinct purposes. Specifically, Linstance improves instance-wise segmentation quality, Lcenter enhances segmentation quality of small instances, Lfalse minimizes the rate of false and missed detections across varied instance sizes, and Lproximity improves detection quality by pulling predicted instances towards the ground truth instances. Through the task of segmenting white matter hyperintensities (WMH) in brain MRI, we benchmarked our proposed instance loss functions, both individually and in combination via an ensemble inference models approach, against traditional pixel-level loss functions. Data were sourced from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the WMH Segmentation Challenge datasets, which exhibit significant variation in WMH instance sizes. Empirical evaluations demonstrate that combining two instance-level loss functions through ensemble inference models outperforms models using other loss function on both the ADNI and WMH Segmentation Challenge datasets for the segmentation and detection of WMH instances. Further, applying these functions to the segmentation of nuclei in histopathology images demonstrated their effectiveness and generalizability beyond WMH, improving performance even in contexts with less severe instance imbalance. •Instance imbalance problem in semantic segmentation is yet to be solved.•Novel loss functions named “instance loss functions” are proposed and investigated.•Instance loss functions consist of Instance, Center, False, and Proximity losses.•Together, they improve instance-level segmentation and detection quality.•Images with varying numbers and sizes of objects will benefit from these new losses.