AI companies develop computer vision behavioral analytics for public spaces, retail

viisights is expanding its biometric behavioral analytics customer pool, as more cities, organizations and partners across the world choose its video intelligence to automatically detect, analyze, and differentiate human behaviors, the company announced.

The company says its Wise software leverages time-based vision analysis and deep learning to identify behavior patterns and tell the difference between a person slipping and falling and being assaulted, or hugging and fighting. The tool can help prevent the spread of COVID-19 by detecting people who do not follow social distancing guidelines.

“viisights unique solution can accurately detect and analyze the behavior of individuals and groups of people using deep learning and time-based vision analysis,” said Asaf Birenzvieg, co-founder and CEO of viisights in the announcement. “viisights behavioral recognition delivers superior video intelligence that can help keep cities and businesses safer and more secure while reducing the manpower and associated costs necessary to monitor vast numbers of live and recorded video sources.”

Whether deployed on-premises or in the cloud, viisights Wise supports integration with multiple cameras and uses video clips instead of images for video streaming analysis. To train the AI engine, it uses convolutional neural networks and LSTM models that generate a unique signature which includes elements such as positioning, movement, size, and relationship. If suspicious or dangerous markers such as fights, riots, vandalism, weapons, blood, loitering, lack of masks, collisions or fire are detected, a notification is sent.

The technology does not identify people or license plates, it only analyzes general behavioral patterns, the company says.

In 2018, viisights was recognized as a Gartner Cool Vendor in AI for Computer Vision.

Last year, the company raised $10 million in Series A funding to invest in behavior recognition technology.

Swiss startup transforms retail sector with behavioral analytics platform

In Switzerland, a local startup called Advertima has secured €15 million (roughly US$17.5 million) Series A funding led by real estate company Fortimo Group, to further develop its behavioral analytics platform for in-store retail, writes TechCrunch. The company is building a machine learning algorithm that it says enhances customer experience in real-time through behavior and demographic analytics.

The “Human Data Layer,” as they call it, features smart inventory management and autonomous checkout, combined with digital signage to deliver smart messages based on user proximity. According to TechCrunch, the platform still needs work in terms of tone and messaging. Advertima claims there is no facial recognition or biometric detection involved in the process, nor does it store videos or images. The process involves “minimal anonymized data.”

“Our system generates a tracking ID every time someone is entering the field of view of a sensor. This ID is a randomly generated hash and is not connected to other tracking IDs, nor can it be connected to other tracking IDs retrospectively,” Advertima told TechCrunch. “If the same person leaves the field of view and returns later, a new tracking ID is generated that is not linked to the previous ID. Tracking IDs are fully anonymous and are kept for analytics purposes only.”

“We do not link tracking IDs to any other data like device IDs or other attributes from external sources,” the company added. “Our system is not designed to and not capable of communicating with other devices like mobile phones. We operate with computer vision only with data privacy in mind to ensure that we deliver a positive experience. From the very beginning, we designed our solution in collaboration with data privacy officials to ensure that we are fully compliant with EU and Swiss data privacy laws, some of the most restrictive data privacy laws in the world.”

Advertima explained that the cashier-less check-out product, if the customer consents to the feature, can track them around the store and share customer ID information with the POS to issue a bill.

Established in 2016, Advertima has collaborated with Swiss retailer Migros and grocery chain SPAR. The technology has been implemented in 14 companies in eight countries.

The investment will help the company expand its technology and improve the platform, while scaling the business across global markets. Over the next two years, Advertima estimates spending some €25 million (approximately $29 million) on building the platform.

“Delivering smart inventory management, autonomous checkout, in-store analytics, and contextual content on smart digital screens will allow grocers and other retailers to maximize the efficiency of their stores, increase their revenues, and generate greater returns per square meter,” wrote in a blog post Advertima Co-Founder and CEO, Iman Nahvi.

“Retailers can actualize an omnichannel strategy to orchestrate better experiences and relationships with their audience. Soon the standard for retailers will be holistically customer-centric: Cashierless checkouts, no lines, individualized experiences, and real-time product recognition for fast, easy, and fun shopping.”

Article Topics

AI  |  behavioral biometrics  |  biometrics  |  computer vision  |  retail biometrics  |  video analytics