PhotoTOC: Automatic Clustering for Browsing Personal Photographs

This paper presents Photo Table Of Contents (PhotoTOC), an interface that helps users find digital photographs in their own collection of hundreds or thousands of photographs. PhotoTOC is a browsing user interface that uses an overview+ detail design. The detail view is a temporally ordered list of all of the user’s photographs. The overview of the user’s collection is automatically generated by an image clustering algorithm, which clusters on the creation time and the color of the photographs. PhotoTOC was developed by design iteration on an earlier clustering user interface: AutoAlbum. PhotoTOC was tested on users’ own photographs against three other browsers: a hierarchical folder browser (with image thumbnails and the user’s own folder structure), a flat detail view with no automatically generated overview, and AutoAlbum. Searching for images with Photo- TOC was subjectively rated easier than all of the other browsers and PhotoTOC’s task performance was not slower than any other browser. This result shows that an automatic organization of personal photographs is effective: it requires no organization effort by the user and yet facilitates efficient and satisfying search.

PDF file


InstitutionMicrosoft Research

Previous Versions

John C. Platt, Mary Czerwinski, and Brent Field. PhotoTOC: Automatic Clustering for Browsing Personal Photographs, Institute of Electrical and Electronics Engineers, Inc., December 2003.

> Publications > PhotoTOC: Automatic Clustering for Browsing Personal Photographs