Extraction and regulation method of melanin and hemoglobin concentration in human face skin image

Abstract

The invention discloses a method for extracting and regulating the concentration of melanin and hemoglobin in human facial skin images, which includes the following steps: obtaining a polarized RGB color image of human facial skin and dividing it into several sub-image blocks; Component analysis and independent component analysis to obtain the mixing matrix set; design the mixing matrix validity determination constraints, filter out the effective mixing matrix subset from the mixing matrix set, and perform statistical operations to obtain the optimal mixing matrix; design the optimal mixing matrix The weight control factor is used to construct a facial skin melanin or hemoglobin concentration distribution image with adjustable concentration and a comprehensive synthetic image. The present invention uses an independent component analysis algorithm and designs the mixing matrix validity determination constraints to solve the problem of unstable concentration extraction results and large errors caused by the interference of the invalid mixing matrix on the concentration extraction in the existing melanin and hemoglobin concentration extraction methods. technical issues.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, and more specifically relates to a method for extracting and regulating the concentration of melanin and hemoglobin in human face skin images based on independent component analysis. Background technique [0002] There are two main types of current skin texture and characteristic analysis methods: contact chemical reagent analysis methods; non-contact digital image analysis methods. The former directly analyzes the skin texture and characteristics from the perspective of ingredients, and usually can obtain very precise quantitative chemical indicators. However, due to its contact operation, it may bring potential harm to the subjects and may cause the subjects to resist. The latter, due to its non-contact characteristics, makes it easier for subjects to accept the analysis process, and has been accepted by more and more skin-related institutions, but the number of quantitative in...

AI Technical Summary

Problems solved by technology

Emphatically, in the present invention, according to the distribution characteristics of melanin and hemoglobin in the R / G / B three-color, the mixing matrix validity determination constraint conditions are designed, which realizes good filtering of the original mixing matrix set, and solves the problem of existing melanin In the hemoglobin concentration extraction method, due to the interference of the concentration extraction by the invalid mixing matrix, the concentration extraction result is unstable and the technical problem is large error

Images