A new bit rate estimation scheme for unsupervised segmentation of video content in IPTV services

In video streaming services limited by network bandwidth such as IPTV services, we present a new bit rate estimation scheme. The proposed scheme estimates the bit rate for each segment in a sequence. To estimate the bit rates, the complexity of each intra-frame is computed as a frame feature. The features are classified as segments using unsupervised clustering. The candidate frames of each segment are used to estimate the bit rate based on a statistical analysis and mathematical model. The proposed scheme can reduce the bit rate while maintaining the target quality in IPTV services.