Histopathology Image Segmentation Using MobileNetV2 based U-net Model
Anusree Kanadath, J. Angel Arul Jothi, Siddhaling Urolagin
Abstract
Histopathology image segmentation is a significant step in the early detection of diseases. Compared to traditional segmentation methods, deep learning models provide better segmentation even though the images are complex in nature. U-net is an effective deep model for the segmentation of medical images. In this work, we propose a MobileNetV2 based U-net model for the segmentation of nuclei regions from Triple Negative Breast Cancer (TNBC) histopathology images. Accuracy, AUC and Jaccard coefficient values are used as the evaluation metrics. The proposed method achieves a Jaccard value of 59% on the TNBC nuclei segmentation dataset.
Topics & Concepts
Jaccard indexSegmentationArtificial intelligenceHistopathologyComputer scienceImage segmentationPattern recognition (psychology)Scale-space segmentationSørensen–Dice coefficientSegmentation-based object categorizationComputer visionPathologyMedicineAI in cancer detectionRadiomics and Machine Learning in Medical ImagingMedical Imaging and Analysis