A method and system for realizing discrete cosine transform based on random calculation

Abstract

The invention discloses a discrete cosine transform realization method and system based on random calculation, wherein the method includes: mapping an image transformation unit (TU) and a discrete cosine transform (DCT) coefficient to a decimal within a preset range, and then The decimals are converted into a random sequence; the random sequence is converted into a random calculation field sequence, and the random calculation multiplier and the random calculation adder are used for calculation; the calculation result is converted into a two's complement form, thereby realizing discrete cosine transformation. By adopting the method and the system disclosed in the present invention, the hardware consumption can be greatly reduced and the working clock frequency of the system can be improved at the same time.

Description

technical field

[0001] The present invention relates to the technical field of digital image compression or video compression, in particular to a method and system for realizing discrete cosine transform based on random calculation. Background technique

[0002] Discrete Cosine Transform (DCT) is widely used in image and video coding standards because of its near-ideal decorrelation performance.

[0003] At present, mainstream image and video coding standards are all block-based coding architectures, which means that the realization of DCT also takes transform block (TU) as the basic unit. Image processing usually uses 8×8 blocks, while video processing is relatively flexible, such as 4×4~8×8 (H.264), 4×4~32×32 (H.265). In the hardware circuit, if the floating-point or fixed-point DCT operation is directly implemented, it will bring more hardware consumption and unacceptable critical path delay. Therefore, the improvement of DCT hardware implementation has two directions: ...

Images