Face2Brain, an Algorithm for Estimating Human State from Video Images.
-Detection of reduced alertness (drowsiness or fatigue, etc.) of drivers and workers using cameras-
ARAYA (Tokyo, Japan; CEO: Ryota Kanai), which develops AI (Artificial Intelligence) technologies and provides solutions, has developed Face2Brain, an algorithm that estimates human states from video-captured facial features by combining AI technology and NeuroTech (BrainTech).
■ Market and background: demand for EEG sensing is growing due to the expansion of the BrainTech market.
Brain tech (neurotech) is a technology that applies "neuroscience," including technology for monitoring brain activity and stimulating the brain for treatment or to improve performance.
The market size of BrainTech is estimated to be around 4 trillion yen (*1) and is expected to increase in the future. There are more than 1,200 companies, mainly in the USA and Europe, and the total amount of investment in companies is expected to increase to USD 8 billion (approximately JPY 900 billion (*2)) by 2020, which is 21 times the amount invested in 2010.
On the other hand, the need to wear an electroencephalograph on the head in order to utilise BrainTech has been a barrier to the practical application of BrainTech due to the burden of wearing the device.
(*1) Brain tech will open up a ¥5 trillion global market.
(*2) Neurotech Investment Digest (Neurotech Analytics Limited)
■ Face2Brain, a system for human state estimation from video images
Face2Brain is an algorithm that estimates brain waves from facial images (patent applied for). By installing this algorithm in devices capable of acquiring facial images (e.g., in-vehicle cameras, PCs with webcams, etc.) or linking it to an API, it is possible to estimate brain waves and detect changes in sleepiness and alertness of workers due to reduced alertness.
Feature 1: EEG can be estimated without the need to wear an electroencephalograph.
Focusing on the link between facial expressions and brain waves, an algorithm has been developed to estimate brain waves from facial images using deep learning, 'without the need to wear an electroencephalograph'. Even in environments where it is difficult to wear an EEG monitor, such as in a car or in a hospital, the brain state can be captured as long as a camera is available.
Feature 2: Capable of estimating arousal levels (e.g., sleepiness) based on brain state.
Brain waves estimated by Face2Brain have a high correlation, especially in the delta wave, which indicates arousal. It is, therefore, suitable for use when driving a car and at construction and manufacturing sites, where attention and judgement are required. It is also envisaged that the system can be applied to safety management, for example, to give rest instructions to workers through monitoring of their condition.
■ Product overview
Product name: Face2Brain
Service website: https: //www.araya.org/service/neurotech/
< Araya Inc. Company profile>
Company name: Araya Inc.
CEO: Ryota Kanai
Location: 24F Ark Mori Building, 1-12-32 Akasaka, Minato-ku, Tokyo 107-6024
Contact form: https://www.araya.org/contact/