Scientists trying to understand the factors underlying facial attractiveness are employing computer technology to specify the essential properties of beauty. They are learning that faces with a lack of distinctness tend to be more appealing, while a symmetrical face does not automatically make someone beautiful. A computer has been trained by Tel Aviv University researchers to identify what people would rate as an attractive female face by automatically extracting measurements of facial features from raw images graded by study participants. The computer condensed the thousands of features it considered, and then used these preferences to predict attractiveness in a new set of faces. Researchers discovered that the computer reflected the human bias for left-left images. Now that the isolation of facial features for analysis has been achieved using computers, University of Western Australia evolutionary psychologist Hanne Lie says the next step will be to reconstruct attractiveness, in keeping with her suspicion that the perception of attractiveness is a holistic process.