NIST logo
*
Bookmark and Share

Video Signal Error Concealment


Objective

To demonstrate to the user community that good quality video reception is attainable for transmission of video over wireless networks and to work with the Institute for Telecommunication Sciences (ITS) to carry out subjective video quality assessment tests and determine the minimum acceptable video transmission quality required to support the NIST Office of Law Enforcement Standards (OLES) public safety programs.


Video Subjective Quality

Wireless video for homeland security applications requires high quality rendition in order to provide a reliable tele-presence assessment of critical situations. While current methods of video compression accelerate transmission by reducing the number of bits to be transmitted, they have the unfortunate tradeoff of increasing signal sensitivity to errors when transmitted over lossy heterogeneous networks. There are two distinct approaches that can be used to improve the video quality in packet loss transmission environments: error concealment and error resilient coding. The error concealment (EC) technique aims at minimizing the deterioration caused by packet loss when the video signal is transported via wireless networks. This later decoding approach makes use of the spatial and/or temporal correlations in the video sequence.

 

 

This figure shows the video input, the encoding, multiplexing and packaging, tranmission, channel (where packet loss may occur), receiving, demultiplexing and depackaging, and decoding to produce the resultant output video


Approach

In this project, a new spatial/temporal interpolation algorithm for error concealment of ITU-T H.264 coded video has been developed for video quality assessment tests. The method aims at interpolating areas in the image, which have been affected by a loss of packet. The project consists of two parts; spatial (intraframe) error concealment (EC), and temporal (interframe) EC. The spatial EC is based on interpolating the missing regions in the frame by using pixels in the same frame. For spatial EC we have first considered bidirectional interpolation. To further improve the performance of the EC, we have developed an edge detection technique to aid the bidirectional interpolation. The edge-detection scheme is based on applying Hough transform to connect separated edge points along the direction of each detected straight line. The lines are used to divide the missing areas into different regions for interpolation.



Contact

Hamid Gharavi
Emerging & Mobile Network Technologies Group
gharavi@nist.gov