I’m building a proof-of-concept layer that intercepts any image or video delivered through a web browser and automatically replaces detected humans or animals with a neutral “shadow” placeholder. The aim is to strip away elements that might provoke irrational responses while keeping the surrounding media unchanged. Here’s what I need you to deliver: • A working browser-based solution (extension, local proxy, or another approach you propose) that processes both static images and streaming video in real time. • Reliable detection of people and animals using modern computer-vision tools (e.g., TensorFlow, OpenCV, YOLO, or alternatives) and seamless substitution with a non-identifying silhouette or blurred overlay. • A simple toggle UI so users can turn the filter on or off without reloading the page. • Clear build/run instructions and concise technical documentation so I can install, test, and demo the concept easily on the latest versions of Chrome and Firefox. I’m flexible on the underlying language or framework; choose whatever stack lets you achieve fast, accurate detection inside the browser environment. Likewise, the integration method is open—feel free to recommend whether an extension, a lightweight proxy, or an injected script will give the best balance of performance, privacy, and maintainability. The solution should start with web images and videos only; we can discuss broader platform support once the core filter is stable. If this sounds like something you can pull off, tell me how you’d approach detection speed, false positives, and resource usage, and let’s get started.