Improving lung cancer pathological hyperspectral diagnosis through cell-level annotation refinement

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: March 8, 2025

Lung cancer remains a major global health challenge, and accurate pathological examination is crucial for early detection. This study aims to enhance hyperspectral image analysis by refining annotations at the cell level creating high-quality dataset of lung tumors. We address challenge coarse manual in datasets, which limit effectiveness deep learning models requiring precise labels training. propose semi-automated annotation refinement method that leverages data diagnosis. Specifically, we employ K-means unsupervised clustering combined with human-guided selection refine into cell-level masks based on spectral features. Our validated using squamous carcinoma containing 65 samples. Experimental results demonstrate our approach improves pixel-level segmentation accuracy from 77.33% 92.52% lower prediction noise. The time required accurately label each slide significantly reduced. While labeling methods an entire can take over 30 mins, requires only about 5 mins. To visualization pathologists, apply conservative post-processing strategy instance segmentation. These highlight addressing challenges improving analysis.

Language: Английский

Histopathology image classification based on semantic correlation clustering domain adaptation

Artificial Intelligence in Medicine, Journal Year: 2025, Volume and Issue: unknown, P. 103110 - 103110

Published: March 1, 2025

Language: Английский