Most forensics investigations today use a copious amount of video data as evidence. However, having to review long video clips with sparse movement can make investigations inefficient, as it’s hard to know where to focus. Investigators can’t afford to miss anything, but they also can’t afford to spend all their time watching videos. Dragonfruit’s video summaries solves this dilemma by squeezing hours-long video into a minutes-long clip, showing only key objects and thereby reducing investigation time. Investigators can search on relevant filters such as clothing or vehicle color, LPR, etc. to narrow down a specific incident. Objects in video summaries can be selected for easy location in the original video.
Dragonfruit uses AI fast-forward to display all the people in an hours-long video in a minutes-long clip. You can filter by people in general or search on their top or bottom clothing color. A time indicator shows how far into the original video the objects are appearing, and a speed indicator shows the speedup factor of the ‘squeeze.’
Dragonfruit allows you to easily filter by vehicles as well. Vehicles can also be filtered by color, and multiple vehicles with different colors can be searched.
The video summaries module works in tandem with other Dragonfruit modules. For example, if you’re using the face mask module, you can create summaries of people with or without masks. Similarly, if you have the LPR module, you can choose a long-running event (e.g. one that lasts for days) and see all the instances of that car within that timeframe by filtering on its license plate. We can also create customized object filters so that you can locate any object of interest.