I have two files for each job: the digital artwork used for printing and a high-resolution scan of the printed sheet. The scan contains twenty repeated copies of the same design, while the original file holds just one. I need a small desktop or command-line tool that will automatically: • align the scan to the original, even if the sheet is slightly rotated or stretched • split or tile the scan so each of the twenty copies is compared against the master design • highlight every deviation—colour shifts, missing or extra elements, mis-registration, smudges, anything that makes the printed piece differ visually from the source • export the marked-up image (overlay or side-by-side) and a concise summary report I can review or forward to our press operator You are free to use Python + OpenCV, C++ with OpenCV, MATLAB, or another well-documented computer-vision stack, as long as the solution runs locally on Windows without complex setup. I will supply sample originals and their corresponding scans for the demo. Before we talk about payment, I must see a working proof of concept on at least one pair of images—nothing fancy, just enough to show the difference-detection pipeline functions accurately. After I confirm the demo meets the brief, we can finalise the hand-off of the full application and arrange immediate payment. When you reply, outline: 1. The libraries and algorithms you plan to use for registration and difference detection 2. How you will handle colour variation thresholds and noise from the scanner 3. Your estimated timeline for the demo and final build I look forward to your proposal.