Binary alpha channel compression for coding of supplementary video streams

When two video sequences are combined in each frame, an alpha channel signal is commonly used. Alpha channels are usually required in professional and studio applications for video editing and post-production processing. However, in other applications related to video broadcasting, alpha channels may be required also at the receiver side. One example is the insertion of sign language interpreters in broadcasted videos to help people with hearing difficulties to follow the programs. The interpreter may be distracting for normal hearing audiences so a solution where the video of the interpreter is switchable would be preferable. In this case, the frame composition, i.e. main program plus interpreter, is done at receiver and therefore the alpha channel needs also to be transmitted. In such an application scenario, it is important to devise efficient alpha channel coding algorithms which require low computational complexity. This paper proposes a lossless encoding algorithm which compresses homogeneous regions with few bits and involving a low amount of computational complexity for alpha channel coding. When compared the HEVC codec in lossless modality the proposed algorithm provides higher compression ratios (up to 54% improvement) and reduced computational complexity. Moreover, the proposed algorithm also outperforms conventional lossless coding techniques such as the Lempel-Ziv-Welch algorithm.