Lung Segmentation in Chest X-ray Images using the Segment Anything Model (SAM)

Abstract

The Segment Anything Model (SAM), introduced by Meta AI in April 2023, represents a cutting-edge tool designed to identify and separate individual objects within images through semantic interpretation. The advanced capabilities of SAM stem from its training on millions of images and masks. Shortly after its release, researchers began evaluating the model’s performance on medical images. With a focus on optimizing work in the healthcare field, this study proposes using SAM to evaluate and analyze X-ray images. To enhance the model’s performance on medical images, a transfer learning approach was employed, specifically through fine-tuning. This adjustment led to a substantial improvement in the evaluation metrics used to assess SAM’s performance compared to the masks provided by the datasets. The results achieved by the model after fine-tuning were satisfactory, demonstrating performance close to that of renowned neural networks for this task, such as U-Net.

Ähnliche Arbeiten