This real time video compression technology decodes and encodes sampled video images using a combination of coefficient-representative and vector-representative signals with supplementary techniques to adapt and optimize them. The two patents have been deemed essential by MPEG LA who have incorporated both into their MPEG-4 Visual Patent Portfolio License Pool: the '834 patent for decoding coded bit planes, and '269 patent for deinterleaving.
The '834 patent describes a method for encoding and decoding an input signal by applying a transform to the input signal; quantizing the transformed signal to obtain a sequence of quantized values represented in the form of ordered binary bits; and for each order of binary bits, encoding a bit plane of the sequence using (run, end) coding to obtain encoded values; and decoding the encoded values to recover the input signal. The patent relates particularly to a method and apparatus for encoding and decoding such signals to achieve bandwidth compression for efficient storage and/or transmission.
The '269 patent uses a deinterleaving step prior to the transformation step in combination with a suitable quantization technique solving the problem of coding images and video with high coding efficiency and relatively low complexity. The strategy uses a Fixed and an Adaptive Deinterleaved Transform Coding approach that develops two types of coder which each approach. One coder operates on an entire picture or VOPs; the other operates on small, local regions allowing for an efficient coding at a range of complex possibilities.