AgileNet: A Rapid and Efficient Breast Lesion Segmentation Method for Medical Image Analysis

Current medical image segmentation approaches have shown promising results in the field of medical image analysis. However, their high computational demands pose significant challenges for resource-constrained medical applications. We propose AgileNet, an efficient breast lesion segmentation that achieves a balance between accuracy and efficiency by leveraging the strengths of both convolutional neural networks and transformers. The proposed Agile block facilitates efficient information exchange by aggregating representations in a cost-effective manner, incorporating both global and local contexts. Through extensive experiments, we demonstrate that AgileNet outperforms state-of-the-art models in terms of accuracy, model size, and throughput when deployed on resource-constrained devices. Our framework offers a promising solution for achieving accurate and efficient medical image segmentation in resource-constrained settings.