SA start-up develops algorithms to determine moods
South African facial recognition start-up Camatica has added ‘mood analytics’ algorithms to its suite of facial recognition artificial intelligence (AI)-powered products.
The AI system allows companies to understand when workers are experiencing challenges by analysing their facial expressions and helping them to come up with solutions to help their employees.
Retailers can also use the system to analyse shoppers who visit their stores. If a shopper shoplifts and they are caught on camera, the system will be able to identify the shoplifter and Camatica will alert the retailer.
The system was developed to provide a strong estimation of what a person’s mood is based on millions of images of people’s different moods.
Johannesburg-based Camatica analyses video feeds from clients’ closed circuit television, connecting it to the video management software and analysing the data in real-time.
The company has developed four products with its AI -powered facial recognition technology.
The company offers Face-Security, where mood analytics can identify people from a list; for example, shoplifters.
Face-Analytics analyses age, gender and mood, and Face-CRM provides the ability to notify the client when a VIP visits the retailer. Face-Attendance is for human resources, to check employees who go on leave without applying for leave days.
Camatica co-founder Laurence Seberini says the intention is to help companies derive value from their video footage using computer vision algorithms.
"For many of us, facial recognition is becoming our first exposure to the growing phenomenon of AI."
Seberini says while artificial intelligence software is good and the algorithms are getting better every day, companies need to be cautious of jumping to conclusions about employees when getting the reports.
“A person might walk by and their face gets caught in a position that might have seemed they were unhappy, and yet they weren’t.”
Seberini advises companies to have a conversation with their stakeholders when deciding to deploy mood analytics, so that invasion of privacy complaints can be avoided.
“I suggest companies incorporate the software into their human resources policy. It’s something that needs to be discussed.”
He says depending on a company’s needs, the facial recognition product price starts at R1 500. “The pricing depends on the company’s modular needs and we bill companies monthly.”
In March, Robert Bravery, manager and senior developer at IQbusiness South Africa, speaking to ITWeb, said SA is missing out on a gem of an opportunity by not maximising the potential of biometric facial recognition technologies in the law enforcement, medical and financial sectors.
Discussing the value of facial recognition and facial detection technologies, Bravery explained that while this is not a new technology, its advancement over the years and ease of use can bring about a revolution in helping cross-sector organisations use the human face to draw a wide range of data, creating new opportunities and lowering productivity time.
"SA is holding back on innovative use of facial recognition technologies, which add much value in creating use-cases that can change the country. Facial recognition augments our current position as human beings and cuts down production times spent on sourcing certain information.”