Method for removing oil gloss of face image

Abstract

The invention relates to a method for removing oil and gloss from a face image. The method comprises the following steps: S1, acquiring an original face image S, and performing skin classification to determine an oil and gloss level; s2, calculating sigma max at each pixel point in the S, and storing the S as a grayscale image I; s3, calculating lambda max at each pixel point of the S, and storing the lambda max as a gray image II; s4, using the gray-scale image II as a guide image, and filtering the gray-scale image I to obtain a preprocessed image; s5, each pixel in the image is preprocessed, and a larger pixel in sigma max is selected; s6, repeating the steps S4 and S5, and executing the next step until each pixel point is detected; s7, processing the preprocessed image to obtain a preprocessed image; and S7, when the oil level of the preprocessed image is lower than S and is not greater than a preset oil level threshold, outputting the preprocessed image as an image D, and otherwise, returning to S2 and updating S. Experiments prove that the method has an obvious oil and gloss removing effect.

Description

technical field [0001] The invention relates to a method for beautifying skin, in particular to a method for removing oiliness from a human face image. Background technique [0002] With the rapid iteration of hardware technology in recent years, whether it is a digital camera or a smart phone, the camera function is continuously improved, the imaging pixels are getting larger and larger, and the definition of the portrait is getting higher and higher. Spots, acne marks, wrinkles and other blemishes will also be captured. To solve this problem, in the field of portrait beautification, it is necessary to perform dermabrasion operations on the human face skin to effectively remove noise, black spots, and blemishes in portrait images, so as to achieve smooth, textured, and softened faces. It is necessary to perform whitening operations on dark and yellowish skin to make the skin fair and rosy. [0003] Existing methods for removing highlights from color images can be roughly ...

AI Technical Summary

Problems solved by technology

Tan et al. proposed the SF map, using the specular-to-diffuse mechanism to successfully separate the diffuse reflection component and the specular reflection component, but the chromaticity of the pixels will change during the separation process, resulting in image color distortion

Shen et al proposed an improved SF map on the basis of Tan, first classifying the image pixels into clusters, and calculating the specular reflection component of the image by using the intensity contrast between the maximum value and the range, but it will also lead to the loss of image texture features

Yang et al. directly applied the low-pass filter, and used the maximum diffuse reflection chromaticity estimation value as the bilateral filtering of the value domain to process the maximum chromaticity map of the image, but due to the lack of global information, the texture characteristics of the image will also be reduced. Kim et al. A MAP optimization framework was proposed to separate the diffuse reflection component of a single image. Although the method has strong stability and good robustness, it is still difficult to preserve texture features when dealing with complex texture images.

Images