Microsoft is proud to be launching our third year of Civic Tech focus at MassChallenge with the Civic Tech Scholarship. The scholarship recognizes 5 startups as they enter the MassChallenge accelerator with a cash grant to help them grow their businesses. Microsoft’s Civic Tech Scholarship aims to identify startups helping elected officials deliver improved services to citizens, increasing communication with residents and enhancing government effectiveness. The ultimate goal is to identify solutions that foster citizen engagement and transparency between government and constituents. This blog post highlights the work of one of the 2016 Civic Tech scholarship recipients.
— Aimee Sprung
Suspect Technologies is interested in building intelligent surveillance analytics and identity verification computer vision products. The company started just a couple years back, seeking to push some of the major advancements in computer vision into the real world. Its first product and the focus of this blog post is automated video redaction for law enforcement agencies.
Body cameras have been introduced rapidly the past few years as a tool of public servant accountability, but making their videos available to the public is complicated and expensive in practice. The FOIA ACT gives US citizens in most locations the right to view most body camera data taken by police officers. However, the content must be properly redacted before released to the public—bystanders and other PII information must be blurred. This turns out to be very hard in practice—it can take several hours to review and redact just an hour-long body camera recording. Video isn’t text—it’s dozens of frames per second, each its own potentially private visual. Redacting personal information out of videos thus is extremely time-consuming and expensive. With body cameras rolling out to most police agencies in the next couple years, law enforcement agencies will spend millions of hours redacting people’s faces and identities in videos. It is critical that the content is reviewed, released, and redacted quickly for the public to view the interaction between it and its law enforcement officers. Redaction thus became a hot bed issue, faster than the public, police agencies, and body camera companies were prepared for.
Suspect Technology’s software solution—-automated video redaction —is optimized for body worn footage, which faces unique challenges due to poor lighting, movement, and low resolution. The AVR system employs an advanced algorithm which automatically tracks faces and identifiers, allowing them to be easily blurred, which reduces the time it takes to redact a video immensely. This product is being rolled out to several thousand law enforcement agencies in the next two years. The software will go along way in helping restore trust in communities.
In summary, redaction is a hot bed issue, one with accountability at its forefront, but with technology and time and money at the root of its problem. It is envisioned that someday with a single click of a button, Suspect Technologies will be able to quickly produce a fully redacted video suitable for review and release to the public. In such a world, trust and accountability would easily be restored to the public. Currently, the company is developing and deploying this and similar technologies at startup accelerator Mass Challenge in Boston. Please connect with them there or at www.suspecttech.com.