Feature Analysis and Classification of Lung and Colon Cancer Images using Advanced Machine Learning Techniques

Classifying lung and colon cancer from histopathological images poses significant challenges

due to intra-class feature similarity, particularly in lung cancer. This research introduces

a Multi-Head Attention-Based CNN (MHAB-CNN) to enhance feature extraction by cap-

turing diverse and localized patterns across dataset folds. Visualization techniques, such

as class activation maps, validated the attention mechanism by highlighting critical regions

influencing classification.

• Supervision