Digital video indexing is a technology that allows to identify and code content in the same way as the human perception does. A video index is a compact digital representation of a video that summarizes the unique characteristics of the material. In the same way as a human DNA is a compact way of recognizing the identity of a single individual, a index is a compact representation of a larger media file. The video indexes are stored in the central database and can later be used to identify video data (files and streams) at different stages of the lifecycle.
Initial video
Fingerprint formation
Fingerprint database
Search
Result
Video for identification
Fingerprint formation
Advantages
Speed
To build digital indexes, we use TAPe technology, which ensures high speed – approximately 0.8ms per frame (on one core of Intel Core 3GHz processor, taking into account Mpeg decoding).
Index size
The list of attributes is compact and takes only 17 bytes per frame. Building indexes for a sequence of frames allows to further reduce the size. For instance, 24 hours of video recorded at 50 frames per second require only 25 MB of space.
Stability
Application of TAPe technology ensures stability against all types of interferences and distortions (inherent in both analog and digital signals), such as various noises, changes in brightness, color, contrast, macroblocking, scaling, shifting, etc.
Conclusion
The developed technology provides compact video indexes resistant to distortions, regardless of the format and type of the original video stream. At the same time, it does not require advanced hardware and allows you to build, compare and search for prints in real time with high accuracy (alpha and beta errors are 0.0165 and 0.0040, respectively) and high speed (it takes about 0.3ms to search for a clip in one hour of video) using standard equipment. This provides a convenient search tool and helps you to effectively manage and protect media assets.
This website uses cookies to ensure you get the best experience