About the project

Spott is multimedia VLAIO project that focuses on enrichment of videos. The main goal of the SPOTT project is to re-identify objects based on detecting the logical story units in different types of video content.

IDLab role

The idea here is to develop a solution that would detect logical story units based on object, and location changes. Based on the features extracted, a distance metric was incorporated to estimate the similarity scores between different frames. The logical story units of the video was calculated by comparing and fusing the similarity scores of different frames. The estimated logical story units were in turn used for performing object re-identification.