In this study, researchers developed an open-source AI model to segment the vastus lateralis and surrounding subcutaneous adipose tissue in order to assess muscle quality and size. A training set of 490 ultrasound images and a validation set of 122 images were obtained from 58 participants using uniform imaging methods. All images were normalized and resized prior to training and validation. A U-Net segmentation model with a ResNet50 backbone was trained over 40 epochs using the training set images. For comparison, segmentation was manually performed on the ground-truth images at the leftmost, middle, and rightmost regions of the muscle. The model’s performance was then evaluated on the validation set using the Dice coefficient and Intersection over Union (IoU) for each region of interest. Overall, the model achieved a mean Dice coefficient of 0.9408 and mean IoU of 0.9581 across the muscle and fat regions. The intraclass correlation coefficients (ICCs) were 0.987 and 0.993 for cross-sectional analysis and average echo intensity respectively, indicating strong agreement between AI-driven segmentation and manual analysis. The open-source model has been released as a user-friendly interface for use in research settings.