Editor’s Note: SSI has partnered with Parks Associates for the creation of DIY FYI, a column designed to help dealers keep track of important smart home market developments, what the competition is and whether they want to jump into something they see as a new opportunity.
Over the past several decades, video analytics technologies have expanded across industry verticals and into consumer’s day-to-day lives. Applications such as tagging people in photos, crowd management at airports, unlocking a cell phone and recognizing someone through a connected doorbell are becoming more commonplace. Home security and smart home companies have adopted video analytics widely, adding value to their peace of mind and home automation value propositions.
There are three main types of video analytics applied in the security space: fixed algorithm analytics, artificial intelligence learning algorithms and facial recognition systems.
- Fixed algorithm analytics systems search for a specific identifiable behavior, and send an alert when that specific behavior is detected.
- Artificial intelligence learning algorithms systems start as a blank slate, and once they connect with a camera, they learn its environment and the usual activity for that area; alerts are sent on detected anomalies.
- Facial recognition systems detect and attempt to identify a person.
Video analytics play a significant role in residential security video surveillance technology, enabling a reduction of false alarms instances and enhancing the ability of the system to detect suspicious situations.
Parks Associates consumer survey of 10,000 broadband households finds that among the 26% of U.S. broadband households that intend to buy a smart video doorbell, the majority rate artificial intelligence (AI) or advanced analytics capabilities as very important when selecting a specific video doorbell to purchase.
Some AI or advanced video analytics features include detecting the presence of strangers or motor vehicles, sending alerts when an object has been moved or removed from view, and identifying security and safety-related sounds, like glass breaking or a fire alarm.
According to Parks Associates research, the anxieties raised by COVID-19 and concurrent social unrest, resulted in many people yearning for peace of mind in the comfort of their home. The firm reports that 33% of smart home device owners report increased usage during the COVID-19 pandemic and that 63% of security system owners plan to purchase a smart home device in the next 12 months. Companies announced new products and enhanced applications starting in the beginning of the year.
At the beginning of 2021, Alarm.com and Arlo launched touchless video doorbells to ensure their consumer’s health and safety. Alarm.com’s touchless video doorbell utilizes internal video analytics to detect whenever a person stands on the doormat instead of motion detection. It immediately triggers a chime, sends a mobile alert to the owner and starts recording visitors. Arlo’s touchless video doorbell uses a proximity sensor to detect approaching visitors. When a person gets within the homeowner’s specified distance of the doorbell, it will give off a chime and emit a light to let the visitor know it has been “rung” and the homeowner has been alerted.
Through the use of video analytics, both companies introduce touchless controls for smart home and security products that eliminate the need for physical contact to spread germs or viruses.
Video analytics can also be used to help reduce false alerts and alarms. Canary, a DIY smart home security provider, launched a professional monitoring service for $9.99 a month in Q1 2021. It features AI-powered video verification to confirm if an emergency is real, helping reduce the occurrence of false alarms. In addition, the video analytics-based security system shares the video with first responders to communicate higher prioritization and execute a faster response time.
Facial recognition systems have made headway in quality and new updated features are steadily progressing in 2021. One of the longest-running facial recognition software involves a photo through a government-controlled database, such as the FBI’s database. Local police departments use a variety of facial recognition software, often purchased from private companies. Ideal performance, however, depends on ideal conditions which do not always exist.
For this reason, real life circumstances tend to reduce the accuracy. Therefore, facial recognition has been known to be wrong when attempting to verify people it was not trained to identify. In response, some counties and states including San Francisco, California, Boston, Portland, Maine, and Portland, Oregon have banned the police from the use of facial recognition. In addition, Amazon just recently extended its ban on sharing their information with police departments.
Smart home device manufacturers increasingly integrate AI features and capabilities into their products to add value across a variety of security-related use cases. While AI applications remain relatively simple today, security system, networked camera and video doorbell manufacturers are experimenting with consumer acceptance of AI-driven features and how best to design user experiences that engender trust, convenience, comfort, peace of mind, and savings.
Over time, with video analytics, homeowners can achieve a more effective, secure and attentive systems in the home that help achieve the real promise of the smart home.
Jennifer Kent, Vice President, Research, Parks Associates & Ariel Taylor, Research Intern, Parks Associates
The post The Burgeoning Implementation of Video Analytics in the Smart Home appeared first on Security Sales & Integration.