Objective:To explore a new method of forensic age detection by analyzing the texture features of human skin surface.Methods:Using computer image processing technique

the forearm skin images of 180 cases of indoor workers in the South West China were processed

and three texture feature parameters were calculated including the number of grids

the mean area of grids and the gray level co-occurrence matrix. The relationship between these parameters and age was analyzed using statistics methods. Results:The experiment data showed that the number of grids decreased with age（male:r=-0.745

P=0.000;female:r=-0.792

P=0.000）. The mean area of grids increased with age（male:r=0.712

P=0.000; female:r=0.809

P=0.000）. The gray correlation in the direction of 0degree increased with age（male:r=0.670

P=0.000;female:r=0.667

P=0.000）. When the proposed method was applied to 20 subjects

the range of error was less than 3 years. Conclusion:Through the quantitative analysis of the skin surface aging characteristic

automatic and effective age estimation are achieved. With the advantages of wide detection range

simple operation and low detection cost

this method has the potential application in clinical practice.