Qognify held a seminar at Secutech 2017 discussing situational management by using big data collected from sources such as video.
Surveillance has always been about having large amount of data. With the trend of big data, it is not just about having these information but how to make the data into useful information.
One of the example Sunny Kong, Sales Director at Qognify, provided was a suspect search system that analyze footage to find a specific person. If a particular suspect is being searched on a recorded footage, there may be 10 to 30 hours of video that needs to be watched. With the suspect search engine, the searched task can be reduced to minutes. “We have a large data pool of videos that has been recorded, now it is put into good use by asking the system to help me to find where this person touch point within a location so we can check his movement and understand what he has done,” said Kong. By having the knowledge of the locations the suspect has appeared, the command center could dispatch law enforcement to each place to increase the chance of catching the suspect.
Kong also mentioned that there are a few different methods to search for a suspect. One of which is using the image from the footage and searching for it throughout the recorded video, telling the system to located the particular suspect base on the highlighted screenshot. Another method is creating an avatar that will give the user the ability to describe the suspect in more detail such as if a person has a backpack or wearing a particular shirt. Lastly, the use of a photograph could also help the system search for the suspect by matching the image.
"With IoT and big data, gone are the days where you keep on recording video into the data box," said Kong.