I have a directory full of media. I would like to search similar videos and hand-remove certain copies (those that are too short, too long, have watermarks, don't have watermarks, &c.).
How can I find videos that are somewhat similar to each other?
I have a directory full of media. I would like to search similar videos and hand-remove certain copies (those that are too short, too long, have watermarks, don't have watermarks, &c.).
How can I find videos that are somewhat similar to each other?
What you are looking for is Digital Video Fingerprinting.
Video fingerprinting is a technique in which software identifies, extracts and then compresses characteristic components of a video, enabling that video to be uniquely identified by its resultant “fingerprint”. Video fingerprinting is technology that has proven itself to be effective at identifying and comparing digital video data.
Here are some relevant companies that support this:
All of these (which are listed on Wikipedia) seem professional paid solutions, some of them do no longer seem to exist. You might be able to find something by searching through project hosting websites...
Most of them decide to keep their algorithms private and rather provide a service than selling a product.
I too have been looking for a program like this. The closest that I have found is Thumbsplus (Cerious.com). It can make thumbnails of most video types. It can then find duplicate or even similar thumbnails of these videos. It is really made for still photos, but it does work for this too.
It helped me.
I found this little program, it's got a trail period, so you can give it a try. Then it cost $50
but it works, uses fingerprinting tech, by making thumbs of videos.
You can try Video Comparer (shareware).
It quickly detects similar videos with image transformations and supports split videos into multiple CDs.
You could also try to fingerprint the audio and find the videos that share a similar audio track.
AcoustId is open source and supports the generation of such audio fingerprints.
You can use fpcalc to generate a fingerprint from the audiotrack. AcoustId provides a small number seach engine called acoustid-index to search for similar fingerprints.
One free and working program (detecting duplicate videos with the same content but different quality and size) we've managed to successfully use was SimilarImages. Although it's discontinued and really old, here are some tips to make it work on newer Windows versions: SimilarImages Reviews - AlternativeTo.net.
a newer addition to this type of tool is i-DeClone, which fares very well against similar programs, both free and paid, to detect duplicate movies hiding in different file formats and resolutions, see for review: https://www.zabkat.com/declone/best-duplicate-movie-files-software.htm#timing