I am looking for a creative solution to the following challenge:
I need a simple video frame manipulation routine, that will embed given
information inside video in a way invisible to the viewer. The data has
to be able to survive through keyframes and recoding with resizing
(to some extent).
The routine is prefrably delivered as ffmpeg libavfilter plugin.
API: [url removed, login to view]
Maximum amount of information that will be embedded: 256 bytes.
In case of questions, please contact me.
I will promptly respond to all queries.
The job got even easier. Below is a link for source code (C/C++) of commonly used invisible watermark algorithms.
I still need a coder that will implement the watermarking as a libavfilter filter.
I have attached a sample filter that puts a logo on video frames. It can be used as a reference for learning frame-level manipulation via the mentioned API.