Using Video Analytics and Sensor Fusion in Law Enforcement: Building a Research Agenda That Includes Business Cases, Privacy and Civil Rights Protections, and Needs for Innovation

Abstract

The panel was structured to reflect four top-level questions: 1. What are the core public safety applications for VA/SF? 2. What are the specific VA/SF tasks needed to carry out those applications? 3. What security, privacy, and civil rights protections are needed? 4. What technology, policy, and educational needs for innovation are most important to address? The panel specified four key business cases for employing VA/SF in public safety, summarized in Figure S.1. The panelists collectively noted that the use of VA/SF to detect crimes and major incidents potentially in progress (accidents, fires) was the highest priority business case. An example comment was that "we want to stop [crime] from happening, not investigate it later." The panel also identified a core set of technical functions for supporting the business cases and needs for core bodies of research on recognizing objects and events in images, video, and other sensor feeds; developing computational infrastructures; and providing a range of security, privacy, and civil rights protections. The body of this report provides detailed lists of common objects and behaviors that VA/SF systems should be able to detect, along with a list of common security, privacy, and civil rights protections that should be integrated into VA/SF implementations. The panel generated 22 high-priority needs for innovation to enhance the effectiveness and security of VA/SF for law enforcement. These 22 needs, combined with discussion about the business cases and enabling research at the workshop, inform creation of an investment roadmap that describes necessary investments and whether they are near-or long-term investments. Table S.1 summarizes the resulting investment roadmap. In general, the panel found that VA/SF were extremely promising technologies for improving public safety. The capability to detect crimes or major incidents was seen as potentially very valuable for society. The panel also said that VA/SF could be of great benefit in investigating crimes and incidents, could provide major time-savers through automatic reporting, and could support performance mon-• There are 22 high-priority needs for innovation to enhance the effectiveness and security of video analytics and sensor fusion (VA/SF) for law enforcement. • VA/SF could be of great benefit in investigating crimes and incidents. • VA/SF could support law enforcement by monitoring officer performance and protecting officers' health and safety. • The risks of VA/SF technologies are significant, with security, privacy, and civil rights protections needed if these technologies are not to be misused or abused. • While VA/SF technologies are indeed promising for supporting public safety, they have a long way to go before reaching their full potential.

Strategic roadmap for interoperable public safety video analytics

“…This functional workflow is not strictly linear; for example, an alert could be generated earlier in the workflow as analytics and edge computing capabilities mature. RAND's report offers a different sample workflow based on a passive monitoring model which is triggered by a positive detection of an object or event by a video analytic, as well as several considerations for the business case for real time video analytics [10]. Regardless of the order, interoperability between each function is essential to achieve the full range of public safety missions.…”

“…The U.S. Department of Justice's research and evaluation agency improves knowledge and understanding of crime and justice issues through multi-disciplinary research and evidence-based knowledge. In 2017, NIJ's sponsored research on the use of video analytics and sensor fusion for law enforcement built upon NISTIR 6184 by identifying four business use cases, innovation needs and a near term investment roadmap [10].…”

“…Whether implemented in support of major events, critical infrastructure protection or through federal grant programs and private partnerships, cities have quickly found value in transitioning video surveillance technology to support localized emergencies and daily local use. In an NIJ sponsored workshop held by RAND, public safety stakeholders identified real-time monitoring to detect crimes and major incidents as the highest priority use case for video analytics and signal fusion for law enforcement [10]. RAND's research details use cases, considerations, and priorities for investment.…”

“…While public safety mission owners seek to expand usage and enable cross jurisdictional access to video data, not all stakeholders are in agreement with the use of video surveillance and the analytics being applied to this data modality, there is push back against video monitoring in some communities [51]. RAND's use cases for video analytics envision a future where computer vision driven alerts at the edge eliminate the need for continuous real time streaming data [10]. From a technology perspective, that future reduces bandwidth and storage requirements.…”

Fostering Innovation to Respond to Top Challenges in Law Enforcement: Proceedings of the National Institute of Justice's 2018 Chiefs' Panel on Priority Law Enforcement Issues and Needs

“…• Digital Evidence and the U.S. Criminal Justice System, which examined needs to acquire and use the rapidly increasing amounts of digital evidence (Goodison, Davis, and Jackson, 2015) • Using Future Internet Technologies to Strengthen Criminal Justice, which examined needs to take advantage of opportunities and mitigate threats resulting from rapidly emerging internet-enabled technologies, including the internet of things (IoT) and the semantic web • Using Future Broadband Communications Technologies to Strengthen Law Enforcement, which examined needs for law enforcement to take advantage of emerging networking technologies effectively (Hollywood, Woods, et al, 2016) • Identifying Law Enforcement Needs for Access to Digital Evidence in Remote Data Centers, which assessed needs for state and local law enforcement to effectively access data needed for investigations but that are stored in remote facilities, cloud services, or social media data centers (Vermeer, Woods, and Jackson, 2018) • Using Social Media and Social Network Analysis in Law Enforcement, which assessed use cases, protections (e.g., security, privacy, and civil rights), and needs for innovation for law enforcement to use social media data and social network analysis methods and tools effectively (Hollywood et al, 2018a) • Using Video Analytics and Sensor Fusion in Law Enforcement, which assessed business cases, protections, and needs for innovation for law enforcement to use video analytics and sensor fusion technologies effectively and safely (Hollywood et al, 2018b).…”