Libraries and code for image cropping with YOLO62n model used in CL-PEFE dataset.

Image Processing These files contain the code and libraries used to create an image segmentation model that was labelled and trained to automatically and accurately identify and separate on a high resolution image a patient's legs and buttocks from the entire background of the image in photographs. The file: 'bg_remove_local (1).py' containts the Python code for the segmentation model The zip file ‘yolo26n_legs_v2.pt.zip’ contains the library used for the image cropping process. The process uses the YOLO26n artificial intelligence model, as YOLO (You Only Look Once) is a family of extremely fast and accurate models for detecting objects in images in real time. The model was pre-trained with the COCO (Common Objects in Context) dataset, providing a good basis for working on object detection, segmentation, and pose estimation tasks. It was then trained to specialise in the recognition of legs and buttocks in photographs captured during 200 epochs with a batch size of 16 to 32 images at a time.With this training, the model achieved high accuracy levels exceeding 98% in the evaluation metrics, reaching a segmentation mAP50 of 0.99 and mAP50-95 of 0.986.