Improved and Automatic Classification of Polyp for Colorectal Cancer

The CRC (Colorectal Cancer) is the lethal disease now a day in all over the world. To detect the polyp and predict about colorectal cancer increased the rate of patient survival. An improved and automated detection system is developed for the detection and classification of polyp. The images based dataset plays a crucial role in the field of medical by applying Deep learning with TL approaches the proposed research implements two models: ShuffelNet and ResNet-50. These models enhanced the accuracy and sort outset the gaps of formal work. The prior work were expensive, time taken, use less size of data and no of classes was also low. There are different parameters applied to achieve the results. Before training models dataset was processed and changed the dimension of all classes in to 224 x 224. Both models were altered according to the problem like the last 3 layers of the models. The accuracy achieved was 98.36 by using above mentioned models respectively. This method indicates the high accuracy which will helpful in the reduction of mortality rate of colorectal cancer.