Colorectal multi-class image classification using deep learning models

Colorectal image classification is a novel application area in medical processing. images are one of the most prevalent malignant tumour disease type world. However, due to complexity histopathological imaging, accurate and effective still needs be addressed. In this work we proposed architecture convolution neural network with deep learning models for multiclass histopathology images. We achieved findings using three models, including vgg16 96.16% modified version Resnet50 97.08%, however Adaptive Resnet152 model generated best accuracy 98.38%. The colorectal dataset publicly available which has 5000 8 classes. study have increased all classes equally, total 15000 been augmentation technique. This consists 60% training 40% testing suggested method paper produced better results than existing categorization methods lowest error rate. For categorization, it straightforward, effective, efficient method. were able attain state-of-the-art outcomes by efficiently utilizing resourced dataset.