Analysis Of The Age Based On The Flexible Forearm Skin Image

Age estimation is more and more widely applied in many fieldsincluding criminal investigation, age-faking in sports and so on. At presentthe main technology used in age estimation is the bone age detectionmethod. Conventional bone age TW3method using radioactive X-rayacquisition wrist bone images can only be used to detect ages of childrenand adolescents, and testing precision of the results is affected by subjectivefactors. According to the development of human skin physiologycharacteristic, skin texture hide the age information. Therefore, this paperdevelops a new age estimation method to take the skin images to extracttexture characteristics and then explore the relationship between age andthese skin parameters. The skin texture age estimation method has thenondestructive, no radiation, detection age range greater advantages.The human skin is affected by many factors such as environment, race,foods, etc, and at the same time the skin of various parts of the body isdeveloped differently, therefore, in order to simplify research model, thispaper is only focus on the skin of healthy volunteers who lived most of theirlives in Chongqing, China and engaged in indoor work in the previous5 years, also this paper just take the forearm skin images of volunteers toexplore the relationship between age and skin texture.The main ideas of this research are as follows:(1) Using digitalmicroscope to get flexible forearm skin texture images;(2) Preprocessingthe flexible forearm skin texture image and extracting the4texture-relatedparameters: angular second order moment (energy), entropy, contrast andrelevant;(3) Statistical analysis of texture parameters changing with age,plotting out the curves between the texture parameters and age according tothe collected data.The result shows that:(1)The skin texture image preprocessinghomomorphic filter can keep in the skin texture image details and achievethe effect of the uniform illumination, at the same time improve the imagecontrast and provides great convenience for the next texture relatedparameters extraction;(2)Further statistical analysis shows that: the male andfemale physiological age has strong negative correlation with the angularsecond order moment (energy), entropy, relevant3flexible forearm textureparameters; male and female physiological age has little correlation with thecontrast of the flexible forearm texture;(3) For male and female skin textureimage,16-level grayscale and32-level grayscale angular second ordermoment (energy), entropy, contrast, relevant have the same change trends.