(сетевой с №3-2015)
(сетевой с №2-2015)
|Articles and journals | Tariffs | Payments | Your profile|
Computational optimization of mutual transformations of color spaces based upon the arithmetic fixed-point.
Abstract.In their article the authors provide their results on systematization of methods for computational optimization of the transformation of color spaces based upon the application of fixed-point arithmetic. The authors formulate the goals and analyze the key problems arising in the situation of computational optimization in the process of color space formation from the standpoint of the speed of operation increase. The principles of transition from a floating point format to a format with a fixed point are stated. The authors also provide an example for the analysis of computational optimization for the mutual transformation of RGB and Y709CbCr. In this article the authros consider the method of computational optimization of the transformation of color spaces based on the application of fixed-point arithmetic. When applying the considered principle of practical implementation, the computation time for an image of 4134x2756 on an Intel Core 2 Duo processor becomes 18 times less. This is a very significant increase in productivity. It is not too difficult to apply this approach to other similar calculations, especially on modern 64-bit and 128-bit processors, when the necessary values fit into a single processor register.
Keywords: RGB, format with a floating point, the format with fixed point, computing optimization, mathematical coprocessor, serial processing of images, parallel processing of images, transformation, color space, image processing
Article was received:26-08-2017
This article written in Russian. You can find full text of article in Russian here .