I have a library of roughly 30,000 product photos that all share the same centre-placed text watermark. I want a robust, repeatable workflow that strips that watermark cleanly, touches up a thin object strip that sits along the right edge, and exports every image on a fixed-size canvas so each product sits in an identical visual position. My preferred stack is Python with OpenCV for the heavy lifting—specifically cv2.inpaint for the watermark and object removal, masking utilities for batch work, and a bit of NumPy to automate the canvas alignment. When in-paint alone leaves artifacts, I usually lean on Photoshop’s content-aware fill / Generative AI, so the script should hand-off problematic frames to a watched folder that Photoshop can sweep through via actions or an ExtendScript. Key points I need you to cover in your solution: • A repeatable Python script that loads the source folder, applies a pre-defined mask for the centre text watermark (same coordinates every time), removes the thin right-hand strip, and inpaints seamlessly. • Automatic re-centering of each product on a fixed canvas so every final file is visually uniform. • Batch speed is crucial—think multi-threaded or chunked processing so thirty thousand files run unattended overnight. • A concise README and commented code so my in-house dev can tweak thresholds later. • Final deliverables: clean JPEG/PNG set with identical naming, the full source script, and the masking templates. I will hand you a small test batch first; once output passes visual inspection and a quick SSIM check, we roll on the full archive. Let me know any additional libraries you need beyond OpenCV so I can clear them with IT before you begin. Must know Python + OpenCV (cv2.inpaint, masking, batch processing) and Photoshop AI/content-aware tools. Needs to handle center watermark removal and right-strip object removal via masks, inpainting, and fixed-canvas output with consistent product alignment.