A novel error detection & concealment technique for videos streamed over error prone channels

In video streaming services and applications, impulse noise occurs due to transmission errors or sometimes it is introduced during signal acquisition. The work presented in this paper proposes a novel impulse noise detection and mitigation (INDAM) method that can significantly recover video frames heavily impaired by impulse noise. The proposed technique uses cyclic redundancy check (CRC) method to create an error mask of the received impaired video frames. This error mask contains pixel-by-pixel error information of the video frames and is exploited further to mitigate the error from the impaired video frame. Each impaired pixel in the video frame is replaced by the average of its corresponding error-free neighboring pixels’ values, hence removing the impaired pixels and replacing them with the newly calculated average. The proposed technique uses the error mask created from the CRC method and uses only those pixels which do not contain error for calculating the averages. Results show that INDAM outperforms other contemporary methods in terms of peak signal to noise ratio (PSNR) and structural similarity index metric (SSIM).