Reading without Words: Analyzing Films through Colors and Images using Indexed Histograms

John Hunter and Dale Hartman (Bucknell University)

This poster displays the past development, present debates, and future possibilities of the Film Search Engine, a searchable database of feature films intended for the use of scholars and the general public. It began by developing a search feature that allows users to find any word or phrase from the closed caption files in the films in the database (sorted by user-chosen criteria). Recently, we have also developed a search feature that allows users to find still images using a color palette (rather than words) as the search criterion, thus allowing the analysis of films using purely a visual means. The poster also discusses the possibilities (and limitations) of histogram oriented gradient searches that would allow users to look for particular objects in the database. A demonstration of the engine will be available, and feedback from conference attendees is warmly solicited.

John Hunter is Associate Professor of Comparative Humanities at Bucknell University.

Dale Hartman is a junior majoring in Computer Science at Bucknell University.

[gview file=”http://budsc16.scholar.bucknell.edu/files/2016/11/Reading-without-Words.pdf”]