e unet deep reinforcement learning Projects | Dual Agent DEEP Q Network | Deep Reinforcement
Intially we load the images from the breast ultrasound images (BUS) dataset and Digital Database for Screening Mammography(DDSM) dataset. The Proposed work encompasses processes such as three stage processing. In this Step the pre-processing of input images removes the noise and improves the image quality providing better results during segmentation and classifictaion respectively. The Steps involed in pre-processing are Hybrid Noise Removal. The acquired images generally contain noises that reduce the quality of images and affect the perfoemance of the method.The Deep Reinforcement learning algorithm called Dual Agent DEEP Q Network is used for fused image segementation. Both the agents are trained effectively to produce accurate segementation results.The Effective un redundant feautres are classified as three severity grades such as grade 1 - Normal tumor, Grade 2 - Moderate Tumor , and Grade 3 Severe Tumor. finally the performance evaluation of this research is performed by considering several metrics which are listed as follows accuracy,precision,recall,F-measure, True Psoitive Rate, False Positive Rate, Computation time.
