TABLE 1.

Algorithm for knee ACL tear mask region extraction with ROI.

Step	Procedure
1	Selection and Loading
	1.1 Extract PD-weighted DICOM images (initial resolution: 512 × 512 pixels) 1.2 Select ACL tear-related slices by expert radiologists
2	Conversion and Initialization
	2.1 Convert DICOM to NII format using Python (maintains 512 × 512 resolution) 2.2 Generate initial black masks in NII format (background value: 0)
3	Annotation with ITK-SNAP
	3.1 Load images and masks into ITK-SNAP 3.2 Annotate ACL tear regions as white by experts (tear regions: value 1) 3.3 Save final masks at 512 × 512 resolution
4	Preprocessing and Augmentation
	4.1 Resize images and masks to 256 × 256 pixels (bilinear interpolation) 4.2 Apply random transformations and ACL tear-specific simulations (flipping, rotation ±1°, brightness ±0.01, contrast ±0.1, tear simulations)