Finding Face Features

We describe a computer program which understands a greyscale image of a face well enough to locate individual face features such as eyes and mouth. The program has two distinct components: modules designed to locate particular face features, usually in a restricted area; and the overall control strategy which activates modules on the basis of the current solution state, and assesses and integrates the results of each module. Our main tool is statististical knowledge obtained by detailed measurements of many example faces. Once an initial location has been estimated, predictions about the positions of other features can be investigated. This can lead to a rapid increase in conndence as other features are identiied in their predicted positions, or alternatively to the initial identiication being quickly rejected as predictions are not connrmed. The program can be tuned either to return an accurate result, or to return fairly probable results very quickly. We describe results when working to high accuracy, in which the aim is to locate 40 pre-speciied feature points chosen for their use in indexing a mugshot database. A variant is presented designed simply to nd eye locations, working at close to video rates. We thank Roly Lishman for many valuable discussions on faces.