The goal of the PERCOL project is to participate to the REPERE multimodal challenge by building a consortium combining different scientific fields (audio, text and video) in order to perform person recognition in video documents. The two main scientific issues addressed by the challenge are firstly multimodal fusion algorithms for automatic person recognition in video broadcast ; and secondly the improvement of information extraction from speech and images thanks to a combine decoding using both modalities to reduce decoding ambiguities. This paper describes the system PERCOLI that participated to the REPERE 2013 challenge and presents the results obtained on the main person recognition tasks.
Index Terms : multimodal fusion, person identification, video processing.