By using Euclidean distance as the classification method, we propose a model to classify humans considering the area of all objects as the feature and palatine rugae as the sample. This model can be applied to victim identification when a wild fire happens. Palatine rugae is resistant to a high temperature and will be intact. On database, the model takes only 1 image for every class and gets the best accuracy at 68.52%. According to the extraction feature model, a result that is robust to image rotation and various distances is provided.

image processing, area feature, human classification, palatine rugae.