Enhancing Deep Learning-Based Colon Cancer Detection Using Attention Module

Cancer is a deadly illness brought on by a confluence of many metabolic anomalies and genetic disorders. Lung and colon cancers are two of the main causes of death and disability in humans. Manual identification of cancer must go through a drawn-out and challenging procedure in analyzing histopathological images. In the field of automatic histopathology image classification, machine learning techniques based on hand-crafted features have gained significant attention. Extraction of features and selection of features is a complex process. Given that deep learning (DL) techniques can automatically extract features from images, DL is a promising option for segmenting and classifying tissues in histopathology images. In this chapter, we have proposed a CNN architecture in fusion with convolution block attention module to identify lung and colon cancer. The proposed CNN model is tested and evaluated using LRC25000 dataset. The model achieves an accuracy rate of 98% using lung classes, 99% using colon classes, and 98.34% using both the classes. The performance of the proposed model is compared with simple CNN without an attention block, and it shows a 3% increment in CNN with an attention block. Thus the proposed CNN with attention module classify lung and colon cancer from histopathology images more efficiently, effectively with fewer parameters.