The project aims to summarize and index a video shot on a GoPro or any mounted camera (an "Egocentric" video) so as to reduce the unwanted monotonous imagery that often take up most of the video time.
When shooting using a GoPro, the video captured are the events during the entire duration whenever the camera is ON. This creates a lot of redundancy wasting a lot of space on the video capturing frames with no information content.
The summarizer reduces this redundant content in the video by cutting out the unwanted parts and only keeping the sections with high information content in the final video.
The indexer clusters frames in buckets to allow for fast image searching. You provide the program with a sample image and it will try to find the image in the video and present with a subsection of the video that contains that image or an image similar to the sample image.
The project is an IntelliJIDEA project and uses OpenCV 3.1.0