Abstract
IPTV services are being developed and provided. To provide these services with sufficient quality, we must properly evaluate and manage users' perceptual quality of service, i.e., Quality of Experience (QoE). To do this, we first proposed a video quality estimation model using the invalid frame. Invalid frame is defined as the total number of degraded frames in video sequences and derived by using video bit-stream information that relates to picture frame type. However, bit-stream information is often encrypted. We then established a packet-layer video quality estimation model, and a picture type estimation method using TS packet header information. Computational cost of the proposed method is high because this method used pattern matching technique. In this paper, we propose a light weight picture type estimation method and verify its validity.
