Scientific journal of Mehmet Akif Ersoy University., Год журнала: 2025, Номер 8(1), С. 35 - 46
Опубликована: Май 22, 2025
Nuclei segmentation in histopathological images is crucial for the processing and analysis of medical images. Manual nuclei challenging due to subjective errors by experts image noise. Before use artificial intelligence analysis, tasks were performed with common classical methods such as thresholding watershed. The development deep learning has led emergence models specifically designed tasks. In this study, LinkNet model supported Vgg16 backbone proposed segmenting CryoNuSeg dataset created nucleus segmentation. After a small number are multiplied data augmentation, feature maps generated using integrated into encoder architecture. results obtained F1 Score, Intersection over Union (IoU), Aggregated Jaccard Index (AJI) values 0.8447, 0.7312, 0.7312 respectively, demonstrate superior performance compared recent studies utilizing same dataset.
Язык: Английский