Automated face annotation for personal photo management

摘要

Automatic face annotation is a process of assigning words or names to human face photos. With the rapid growth of digitized photos captured for our personal collections, album management becomes very cumbersome. Commercial image organizer products such as ACDSee, F-Spot and Picasa allow users to automatically group images by folder, album, and date; thus easing browsing and retrieval. However, a study showed that users usually tend to organize images into collections by date, location, and events. Therefore, we proposed an automated face annotation using face features, clothing colours and date the photo is taken. A total of 175 photographs comprising 138 frontal faces are selected from events such as Eid celebration, birthday celebration and outdoor occasions. The face features are described using eigenfaces, while the clothing colours are represented using La*b* colour histogram. Julian date format is added into the feature vector and metric-distance similarity measure is utilized to recognize the face. A recognition rate of over eighty percent is achieved for our proposed method.