DENSO TEN’s AI predicts near misses and signs of drowsiness

DENSO TEN has developed AI technology that predicts a person’s inner state, such as near misses and signs of drowsiness, from facial images taken with a camera.
First, an AI model is generated that predicts the relationship between a person’s inner state, such as “a near miss, a startle,” or “a dizzy head,” which can be determined from biometric information such as brainwaves and heart rate, and the feature values that represent the position and shape of facial parts such as eyebrows, eyes, nose, and mouth from the facial image at that time.
Next, by extracting feature values from the facial image captured by the camera and comparing them with this AI model, it becomes possible to predict a person’s inner state.
The company plans to use this technology to contribute to a safe and secure mobility society in the future.
※Translating Japanese articles into English with AI