Efficient Cumulative Probability Distribution Estimation for Arithmetic Coding

Cumulative probability estimates of binary symbols are performed by finite state machines selecting binary symbols from the predefined finite set, in order to feed state-of-the-art arithmetic coders in image compression devices. This paper discloses novel method for cumulative probability distribution estimation, based on low pass filtering with varying dominant pole, providing high estimation precision and short code length. This embodiment utilizes only integer arithmetic, without multiplication or division operations, thus providing significant reduction in necessary hardware resources of IoT devices.