ACMU-Net: An Efficient Architecture Based on ConvMixer and Attention Mechanism for Colorectal Polyp Segmentation

Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths worldwide, and early detection plays a crucial role in improving patient outcomes. Colonoscopy is the gold standard for detecting CRC, but manual analysis of colonoscopy images is labour-intensive, subject to human error, and often inconsistent. To address these challenges, we propose ACMU-Net, a novel deep learning model designed to enhance the accuracy and efficiency of polyp segmentation in colonoscopy images. ACMU-Net combines the strengths of U-N et for medical image segmentation, ConvMixer for effective feature extraction, and CBAM (Convolutional Block Attention Module) for refined attention mechanisms, resulting in a robust architecture tailored to this task. Our model outperforms existing approaches by achieving higher segmentation accuracy while significantly reducing the number of parameters, making it computationally efficient and suitable for real-time applications. ACMU-Net demonstrates strong potential as an advanced tool for automated polyp detection, contributing to improved CRC diagnosis and treatment outcomes.