Functional model of searching for object in a video stream

Existing video data processing systems use lossy compression methods due to the transmission of reference (I-frames) compressed by spatial coordinates, as well as reference (P- and B-frames) compressed by spatial-temporal coordinates, the number of which determines the degree of compression of the video sequence. At the same time, the peculiarities of the perception of this type of data by a person are taken into account. To date, the problem of eliminating the redundancy of reference frames has been solved to a certain extent, whereas this problem is relevant for reference frames. The basis for encoding reference frames is the construction of motion vectors of individual pixels or groups of them. Also, the efficiency of algorithms for eliminating time redundancy is improved by increasing the efficiency of encoding algorithms by transformation and interpolation.

Hali tarjima qilinmagan