Efficient Knowledge Distillation of SAM for Medical Image Segmentation

TL;DR

This paper introduces KD SAM, a knowledge distillation method that reduces SAM's computational demands for medical image segmentation while maintaining high accuracy, enabling real-time applications in resource-limited settings.

Contribution

The paper presents a novel dual-loss knowledge distillation approach for SAM, improving efficiency without sacrificing segmentation performance in medical imaging.

Findings

KD SAM achieves comparable or better accuracy than baseline models.

Significantly fewer parameters in KD SAM enable real-time deployment.

Effective balance between segmentation quality and computational efficiency.

Abstract

The Segment Anything Model (SAM) has set a new standard in interactive image segmentation, offering robust performance across various tasks. However, its significant computational requirements limit its deployment in real-time or resource-constrained environments. To address these challenges, we propose a novel knowledge distillation approach, KD SAM, which incorporates both encoder and decoder optimization through a combination of Mean Squared Error (MSE) and Perceptual Loss. This dual-loss framework captures structural and semantic features, enabling the student model to maintain high segmentation accuracy while reducing computational complexity. Based on the model evaluation on datasets, including Kvasir-SEG, ISIC 2017, Fetal Head Ultrasound, and Breast Ultrasound, we demonstrate that KD SAM achieves comparable or superior performance to the baseline models, with significantly fewer parameters. KD SAM effectively balances segmentation accuracy and computational efficiency, making it well-suited for real-time medical image segmentation applications in resource-constrained environments.