Generally speaking, directly downloaded unwatermarked videos usually contain the following key information:

1. The MD5 of the original video;
2. The video fingerprint information embedded in the meta information of the original video on social media platforms. [Refer to the table below]
3. Hidden watermark formed by other video algorithms, also known as hidden watermark.

Video fingerprint is the most basic element of duplicate detection, and the basic operation of video de-duplication is to remove these video fingerprint information and do some micro editing in the video.

Why is video deduplication necessary when we upload same video to TikTok?