Andrew Ng, co-founder of Google Brain and Coursera, describes Artificial Intelligence (AI) as the “new electricity,” drawing a parallel between the profound impact both technologies have had on daily life: In the same way that electricity evolved, becoming a part of everyday home and business functionality, Artificial Intelligence is revolutionizing the ways we live and work.
Extracting and leveraging volumes of data, machine learning techniques enable technology to identify and analyze information and transform it into actionable intelligence.
For example, many homes, businesses and government agencies rely on video surveillance for security purposes. Traditionally, video has been used as evidence in police investigations or for monitoring for safety threats or criminal activity. But relying on video surveillance can be inefficient, because human operators cannot actively focus on watching video for hours at a time. An average investigation depends upon the review of thousands of hours of video, and efficient and effective video review is not possible without the help of technology.
Today, video content analytics (VCA), driven by Deep Learning and Artificial Intelligence, help organizations accelerate investigations, attain situational awareness and maximize their investments in video surveillance beyond security purposes, enabling them to derive operational intelligence and uncover patterns and trends.
From in-store retail to urban planning and law enforcement, video analytics enables the detection, identification, classification and analysis of video objects and behaviors, empowering organizations to productively review video, proactively respond to events and predictively plan ahead.
Today, it’s estimated that 85% of all criminal cases are solved using some sort of digital evidence, sourced from social media, email, mobile phones, the cloud and security surveillance. Video evidence has always been critical for law enforcement investigations, but video analytics and face recognition have made it possible to more effectively and efficiently pinpoint persons of interest and identify anomalous behavior.
With face recognition, law enforcement can track specific suspects across video surveillance resources. By configuring alerting rules based on face recognition and suspect watchlists, police can receive real-time alerts when persons of interest are detected by the video content analytics system. It can also be used post-event to search video for appearances of specific suspicious persons, helping police find evidence they need faster.
When no suspect has been identified, law enforcement can utilize video analytics to detect and extract objects that appear in video, identify and classify them and then index the data. This enables video to be searched and filtered based on known criteria; configured to trigger alerts when specific behaviors and conditions are met; and aggregated over time to uncover actionable intelligence about surveilled trends and patterns.
Today, retailers typically utilize video surveillance for loss prevention, crime deterrence, and employee oversight. With the introduction of AI-based video content analytics, retailers can extend the value of video beyond traditional security applications.
Measuring store hotspots, shopper traffic flows and dwell patterns, and product display activity, retailers can identify trends and evaluable them over time for optimizing store layout and driving conversions: identifying underutilized space, uncovering where and why bottlenecks form, and quickly recognizing in-store behavioral patterns.
Much like retailers, cities are also interested in understanding movement patterns, with traffic optimization being a major focus of city governments and law enforcement. With security surveillance and video analytics providing full insight into pedestrian and vehicular activity, cities can leverage data intelligence to optimize pedestrian and vehicular traffic flows.
Driven by artificial intelligence and extracted data, video analytics solutions enable cities to quantify and classify vehicles and pedestrians, discover their movement patterns and identify traffic hotspots. Cities can use this information to identify the causes of bottlenecks and how to prevent them; understand where public transportation is inefficient or insufficient and optimize transit offerings; and develop future infrastructure by understanding where traffic lights, parking lots and traffic circles are needed.
In cities where alternate forms of transportation are popular, such as bikes and scooters, law enforcement can leverage video analytics technology to keep riders, pedestrians and drivers safer: By making it easier for law enforcement to identify traffic violations – such as riding without a helmet or in areas not intended for bike or scooter traffic – police can better respond and enforce safety policies. With comprehensive insight into movement and navigation patterns, cities can support traffic optimization, enforcement and safety.
Utilizing Deep Learning techniques to transform video into a searchable, quantifiable and actionable asset for leading enterprises and law enforcement agencies worldwide, video analytics is just one aspect of the major technological revolution powered by AI. As artificial intelligence continues to develop and become more sophisticated, the advances for every industry become more significant, sparking further innovation for everyday impact.