Scientists at RIT have embarked on a million-dollar project aimed at giving insights into the day-to-day activities of police officers. The initiative is anchored on machine learning techniques developed by the researchers to analyze body-worn camera footage, with the goal of flagging relevant videos and enabling the public to better understand the challenges faced by police officers in their line of work.
According to the RIT scientists, the Rochester Police Department (RPD) produces about 70,000 hours of footage every day, which is too much for officers or supervisors to review entirely. The use of machine learning algorithms will help monitor the footage and flag specific words or identify biased behaviors by RPD officers. Additionally, the algorithms will detect effective de-escalation tactics used by officers, and transcribe body-worn camera audio.
John McCluskey, an RIT scientist who has been working on the project for three years, said the technology would enable them to examine cases that start at similar points but wind up with different outcomes. “We want to ultimately have people code videos that have the beginning that is very similar but generate different outcomes to figure out what branches conversations might take or physical actions might take that generate outcomes that are deescalated versus escalated,” McCluskey explained.
The RIT project is expected to provide insights into police officer’s day-to-day work, which is often under the public’s scrutiny. It will also provide an opportunity for researchers to understand the nuances of police interactions with the public, which could ultimately help in improving police-community relations.
FingerLakes1.com is the region’s leading all-digital news publication. The company was founded in 1998 and has been keeping residents informed for more than two decades. Have a lead? Send it to [email protected]