engleski

Influence of Segmentation Parameters on Video Quality in Dynamic Adaptive Streaming

Dynamic adaptive streaming has been evolving during the last decade in order to meet high expectations regarding video quality of streamed video sequences on various devices. Every year various dynamic adaptation algorithms that try to improve user quality of experience are developed. Another important aspect of adaptive streaming is the video segmentation process as well as the dataset used for testing purposes. The goal of the research presented in this paper was to improve the segmentation process. Structural Similarity Index (SSIM) was used to determine coding bitrates where the overlapping of video quality for different spatial resolutions occur in order to specify proper representations for the segmentation. Two video sequences with different spatial and temporal activity were encoded based on selected representations and tested using the Basic adaptation algorithm (BAA) and Segment-Aware Rate Adaptation (SARA). Compared to previously available segmentation, test results show improvement in SSIM values especially for a video sequence with high spatial and temporal activity.

video segmentation, adaptive streaming, SSIM, bitrate selection

nije evidentirano