There is a general expectation from local Law Enforcement teams in India to have a more efficient tools to Investigate incident with the help of video recording registered through video cameras installed across the capital city and state. Currently, for any investigation to happen, teams has to manually go through various video recordings to identify clues and it is often a laborious task resulting in longer time to complete investigation and in most of the cases, due to other higher priority issues, certain cases are not attended to and as time goes by, critical video recording a purged due to want to additional storage space.

In such scenarios, it will be often beneficial if there exist a tool that can infer video streams to identify objects (vehicles, humans etc.,) and if such objects can be converted in to a metadata format which are often in textual form, then such information can be used in future to analyze pattern of incidents in an efficient way. Metadata is used to describe a scene or a frame in a video stream.

Ceyark video inference (CVI) Platform offers a DNN model that performs objects classification and detection on video streams either at real time or in batch mode and convert object information into metadata objects. These metadata objects are then transformed to messages and will be stored in Datastore. When law enforcement team has a need to go back and understand a situation for which they need a starting point, they can start the incident analysis through a search query on the datastore by passing known parameters like time, objects type, geography plus any other known parameters.

If there exist one or more metadata records that match a search criteria, then one can use the metadata information to identify the corresponding video file that contain the actual video that can assist further in the investigation. But it should be noted that if the corresponding video records are purged that relate to the metadata record, then metadata information can only be used for further leads and can not be used in the court of law.

Do reach out to us to get more details about our platform offerings and how we can work together to meet your needs.