Early Detection of Colon Cancer using Fine Tuned Light Weight CNN Model

There are around 1.8 million new cases of colorectal cancer diagnosed every year, making it one of the most prevalent forms of the disease. Therefore, the infection and mortality rates can be reduced with early identification of this cancer. Most diagnostic colonoscopy systems are currently incorporating artificial intelligence approaches validated for predicting advanced cancers. It is common practice to combine convolutional neural network-based patterns with image patches and preprocesses. The purpose of this paper was to develop a novel lightweight deep learning Convolutional Neural Network (CNN) that uses Adam optimizer for reliable colon cancer detection. Applying a database of publicly available histopathology images, the proposed approach is analyzed in the context of the existing best practices for colon cancer detection analysis. Based on the results, it is clear that the proposed deep neural network for colon cancer diagnosis achieves an accuracy of 98.40 percent, which is the highest accuracy achieved by any of the existing deep learning methods.