Interactive multimodal video search: an extended postevaluation for the VBS 2022 competition

CLIPbased texttoimage retrieval has proven to be very effective at the interactive video retrieval competition Video Browser Showdown 2022, where all three topscoring teams had implemented a variant of a CLIP model in their system. Since the performance of these three systems was quite close, this postevaluation was designed to get better insights on the differences of the systems and compare the CLIPbased textquery retrieval engines by introducing slight modifications to the original competition settings. An extended analysis of the overall results and the retrieval performance of all systems’ functionalities shows that a strong text retrieval model certainly helps, but has to be coupled with extensive browsing capabilities and other querymodalities to consistently solve knownitemsearch tasks in a largescale video database.