Heavy video unique identifier for Win11

Desktop Apps, Python — incorrectly specified categories?

135 USD

Project translated automatically. Log in or register, to view the original

Need software for win11 that can batch generate dozens to hundreds of unique versions of videos without quality loss, bypassing modern duplicate detection systems (video hashes, neural network embeddings, audio fingerprinting).

The task is not simple and involves work on several levels:
geometric transformations (including perspective warp),
color and light (LUT, contrast, local changes),
textural changes (animated grain, SR → downscale),
temporal transformations (interpolation, motion blur, time-warp),
light AI/ML effects (super-resolution, style, depth effects),
adaptive audio processing (if present),
randomization of parameters and generation of metadata “device”.

All changes must be imperceptible to the eye but strong enough for the videos to be considered different by algorithms.
Very long encoding is allowed, especially in Hard mode (1–2+ minutes per 1 second of video).
A careful architecture is required with one final re-encoding and fully local operation.
You can write from scratch or modify an existing script. Any programming languages, libraries, neural networks, etc. are acceptable - the main thing is the result.

Informal part:
Perhaps you would be interested in writing something that will outplay the best algorithms on the planet every day, created by the highest-paid engineers. The top league :)

Please do not disturb vibe coders — trust me, you will waste your time, this is a task of a different level.