High-efficiency facial beautification device easy for parallel realization

Abstract

The invention discloses a high-efficiency facial beautification device easy for parallel realization, and the device comprises an image obtaining module, an image overall beautifying module, an image enhancement processing module, a skin region shading processing module, and an image fusion module. The image obtaining module is used for obtaining an RGB color image, and the output end of the image obtaining module is connected with the image overall beautifying module, the image enhancement processing module and the skin region shading processing module. The output end of the image overall beautifying module, the output end of the image enhancement processing module and the output end of the skin region shading processing module are connected with the image fusion module. The device achieves the processing of a plurality of sub-modules of a system with the help of the hardware acceleration characteristics of a GPU, and can solve a problem that the efficiency is low when a CPU is in use. Meanwhile, the plurality of sub-modules proposed and used by the device can be well applied to the GPU acceleration processing, achieve instant effect display while guaranteeing the instantaneity of efficiency, and are used for the special effect processing of a single image.

Description

technical field [0001] The invention relates to a high-efficiency and easy-to-implement cosmetic beauty device. Background technique [0002] The beauty method usually consists of multiple basic steps, including the removal of noise such as external light and human skin in the image, human skin detection, face detection, freckle removal, skin whitening, image fusion, etc. [0003] As the most basic and important part of image denoising, it plays a vital role in subsequent algorithm processing. At present, there are many denoising algorithms, usually including Gaussian smoothing, bilinear filter smoothing, average filter smoothing, and block-based smoothing. Algorithms such as matching 3D denoising, in which the performance and effect of each algorithm have different degrees of difference, and each has different degrees of limitations, which has a greater impact on the selection of algorithms for application scenarios. For example, the average filter smoothing efficiency is ...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to overcome the deficiencies of the prior art, to provide a high-efficiency and easy-to-implement parallel beautification device, with the help of the hardware acceleration characteristics of the GPU, the problem of low efficiency when using the CPU can be solved, and at the same time, each subsystem proposed and used by the system The modules can be well applied to GPU accelerated processing, and real-time effects can be obtained while ensuring real-time efficiency

Images