Colon Cancer Detection Model using Multiscale Feature Fusion

Colon cancer remains one of the deadliest malignancies worldwide, necessitating early and accurate detection for improved patient outcomes. This study explores deep learning-based classification models for colon cancer detection in medical imaging, with a particular focus on evaluating the performance of the AlexNet architecture. A comparative analysis was conducted against VGG, EfficientNet, and ShuffleNet to determine the most effective model. The results demonstrate that AlexNet achieved a validation accuracy of 98%, an F1-score of 0.98, outperforming alternative models. The validation accuracy of other models was VGG (97%), EfficientNet (32%), and ShuffleNet (96%). The findings highlight the efficiency of AlexNet, making it a promising solution for real-time medical diagnostics. Its lightweight architecture ensures efficient feature extraction while maintaining high classification accuracy, offering a computationally efficient and reliable approach to automated colon cancer detection.