MAMMOGRAPHY IMAGE BREAST CANCER DETECTION USING DEEP TRANSFER LEARNING

All over the world, the bosom disease is analysed in about 12% of women during their lifetime and is the driving motive after the death of women. Since early findings can improve treatment outcomes and have time for patients with prolonged endurance disease, it is noteworthy to construct bosom malignant growth detection procedures. The Conversational Neural Network (CNN) can naturally remove highlights from images and order them later. Massive marked images may be required to train CNN without any preparation, which is possible for certain types of clinical picture information, A promising mechanism is to implement transfer learning on CNN. In this article, we applied the MIAS dataset on three training strategies: CNN to highlight previously prepared VGG-16 models with input mammograms, and to create a Neural Network classifier. Used these highlights for and refreshed the load. By back-spreading (tweaking) to identify abnormal areas in pre-prepared VGG-16 model layers. Compare proposed models and parameters related to performance metrics.