A novel cell contour-based instance segmentation model and its applications in HER2 breast cancer discrimination
Abstract: Highlights•We propose B-channel Topology Analysis (BTA) based on boundary tracking to determine the initial cell contour.•We propose Adaptive LSTM Alignment (ALA) to utilize the local contour information more accurately.•We propose Weighted Loss (WL) to train the HER2 Breast Cancer samples segmentation model.•We propose a HER2 Breast Cancer Discrimination method to assist in sample statistics.
Loading