Personal identification is the most important process in forensic works. But in worst case scenario as natural disasters or major accidents that found only mutilated body with damaged tissues and bones or cannot examine by autopsy from religious problems. The sex identification and height estimation can be made from the distance of proximal epiphysis

This study aims to examine the relationship between gender and height to the size of the 15 length from 6 marking point in proximal epiphysis of femur, and develop regression model to be used in Thai population for gender and height forecasting. Data were collected from 116 radiographic picture from femur AP stance of medical patients in police general hospital. Then calculated by statistical methods using Logistic Regression to find the correlation with gender, and Linear Regression correlation with height. But many lengths of proximal epiphysis falsely show insignificance to sex and height due to colinearity of data. Then using the Stepwise method to fine the variables in regression model that correlated most with sex and least standard error on height estimation model for each sex. The results showed shortest distance of curvature change point on neck of femur and range from lateral side of curvature change point on neck of femur to medial side of maximum distance on head of femur are statistically significant with sex, with 84.5% of accuracy from model. But only 10.1% male and 19.4% female height can be estimated from the model. Distance on femoral epiphysis is good for sex determination but could not apply to estimate height.