I have a collection of elephant photographs that sit in the medium-resolution range—some are perfectly usable, others suffer a touch of blur. For every image, I need clean, separate crops of the left and right ear produced automatically by software rather than by hand. Here’s the flow I have in mind: you build or adapt a computer-vision routine (Python + OpenCV or any framework you prefer) that detects both ears in each photo, isolates them, and saves each ear as an individual image file while preserving as much native resolution as possible. A short script or command-line tool that runs in batch on my local machine is ideal; clear instructions for installation and execution are part of the hand-off. Deliverables • Source code with comments • Read-me or quick-start guide • Folder of sample outputs that match the original filenames (e.g., IMG_123_left.jpg, IMG_123_right.jpg) Accuracy on slightly blurry shots is important, so please outline any pre-processing or model training you plan to apply to keep detection consistent.