Breast Cancer Lesion Detection and Segmentation Based On Mask R-CNN
Hama Soltani, Mohamed Amroune, Issam Bendib, Mohamed Yassine Haouam
Abstract
Breast cancer is an obsession that haunts all women. but early detection for it increases the cure rate, for attain this objective It is very important to create a system to diagnose suspicious masses. On the other hand, is a difficult task due to the fact that breast lumps vary in size and appearance. In this paper, we propose an automatic breast mass segmentation method based on the Mask RCNN model of deep learning using detectron2.our model is trained and testing using the public dataset INbreast. The proposed method achieved results with precision and F1 score 95.87 and 81.05 on INbreast dataset, respectively.
Topics & Concepts
SegmentationArtificial intelligenceComputer sciencePattern recognition (psychology)Breast cancerDeep learningImage segmentationTask (project management)CancerMedicineInternal medicineManagementEconomicsAI in cancer detectionBrain Tumor Detection and ClassificationRadiomics and Machine Learning in Medical Imaging