Real Time Efficient Foreground Extraction with Video Processing

Video analysis is large source of unstructured information today. Introduction to HEVC standard has tremendous potential to lower the cost of data transfer and storage. HEVC provides forty percent or more than that of reduction in bit rate at the same visual quality. Looking over problem of limited bandwidth it became necessary to compress video data. To compress the data spatial and de correlated data is removed. This causes compacted information to fill in as a key asset for machine learning with altogether less examples for preparing. In this venture, a great way to deal with closer view extraction/division is expected abuse novel spatio-transient de-related piece alternatives taken straightforwardly through HEVC compressed video. Various techniques, in distinction, work on pictures which are uncompressed providing vital storage and machine resources not just for the decryption method before data formatting. The planned approach has been quantitatively and qualitatively evaluated against Gaussian mixture model.