Back

MedSegBench: A Comprehensive Benchmark for Medical Image Segmentation in Diverse Data Modalities

Kus, Z.; Aydin, M.

2024-08-28 health informatics
10.1101/2024.08.26.24312619 medRxiv
Show abstract

MedSegBench is a comprehensive benchmark designed to evaluate deep learning models for medical image segmentation across a wide range of modalities.. This benchmark includes 35 datasets with over 60,000 images, covering modalities such as ultrasound, MRI, and X-ray. It addresses challenges in medical imaging, such as variability in image quality and dataset imbalances, by providing standardized datasets with train/validation/test splits. The benchmark supports binary and multi-class segmentation tasks with up to 19 classes. Evaluations are conducted using the U-Net architecture with various encoder/decoder networks, including ResNets, EfficientNet, and DenseNet, to evaluate model performance. MedSegBench serves as a valuable resource for developing robust and flexible segmentation algorithms. It allows for fair comparisons across different models and promotes the development of universal models for medical tasks. The datasets and source code are publicly available, encouraging further research and development in medical image analysis.

Matching journals

The top 1 journal accounts for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.