According to one embodiment, an indexing module is configured to analyze a plurality of still images and to generate index information indicative of attributes of the plurality of still images. The index information comprises recognition results of a plurality of face images in the plurality of still images. An extracting module is configured to extract first still images and second still images from the plurality of still images based on the index information. The first still images include face images of a selected person, the second still images are relevant to the selected person, and the second still images comprise at least a still image including a face image of another person relevant to the selected person. A player module is configured to play back a movie using the extracted first still images and the extracted second still images.