Coresets for Autonomous Vehicle Video Summarization and Search With recent advancements in mobile technology our day to day lives are being recorded in ways never thought possible before. One of those ways is through the vast amount of video due to the high quality cameras on cell phones. This project aims to tackle the challenge of navigating a massive library of content with a textual based search. By using a k-segment mean coresets algorithm we can efficiently summarize this large data set and use object detection on the key frames in order to map sections of this library to a textual representation which will allow for easy searching. This SuperUROP aims to provide a web interface allowing users to upload content and search through their video libraries with text based queries.
I am participating in the SuperUROP program because it will give me the opportunity to specialize in a field I am interested in. I am very excited to work on a new project with the Lab I have been apart of since Freshman summer and look forward to learning more about computer vision and machine learning.