A fully automated quantitative analysis method based on deep learning algorithms for immunohistochemical staining expression intensities
Yongjian Deng, Bojun Cai, Xiaomei Wang · Intelligent Oncology · 2025
The primary focus of this paper is to explore the application of deep learning techniques and image processing algorithms in immunohistochemistry analysis, specifically targeting automated quantitative methods for nuclear, membrane, and cytoplasmic expressions of animal cells in whole-slide images. By employing optical density separation techniques to differentiate between Hematoxylin and 3,3’-diaminobenzidine staining components, combined with the CellViT nuclear segmentation algorithm and the region growing algorithm, precise identification and quantification of cell nuclei, membranes, and cytoplasm are achieved. Experimental validation demonstrates that the proposed algorithm exhibits excellent performance in terms of accuracy and recall. Compared to traditional manual interpretation, this algorithm demonstrates higher accuracy in specific quantitative metrics.