Artificial Intelligence Driven Colorectal Cancer Classification via Deep Learning Technique

Abstract

Collateral cancer is a serious concern for patients with primary tumors, as the development of secondary tumours can significantly reduce survival rates and increase the complexity of treatment. Colon polyps are a common precursor to colorectal cancer, and early detection is critical for successful treatment. Traditional methods of polyp detection include colonoscopy and biopsy, which can be invasive and time-consuming. Machine learning algorithms have shown promise in detecting polyps in colonoscopy images, and in this study, we explore the use of convolutional neural network to detect polyps in colonoscopy images. Early detection and prevention of collateral cancer are crucial for improving patient outcomes, but accurately predicting the risk of secondary tumors can be challenging due to the complexity of cancer progression and the multitude of factors that can contribute to tumour development. Machine learning algorithms have shown promise in predicting cancer outcomes based on patient data, and in this study, we explore the use of a convolutional neural network (CNN) to predict the likelihood of collateral cancer in patients with primary tumors. The study uses a dataset of colonoscopy images, including both positive and negative cases of polyps. A CNN model is developed using this dataset to classify images as either positive or negative for polyps. The model is trained using a supervised learning approach, where the network learns from labelled examples of images with and without polyps. The accuracy of the CNN model is compared to other polyp detection methods, such as traditional image analysis techniques and other machine learning algorithms.