Face Orientation Estimation Using Depth-Gyro Sensor

Shibaura Institute of Technology (Japan)

Researchers at Japan’s Shibaura Institute of Technology (SIT) used a depth-gyro sensor to collect point cloud data for various face orientations in order to train a deep learning-based classification model in face orientation estimation. The researchers altered a face’s horizontal angle relative to the camera from more than 90 degrees to less than 90 degrees using step sizes of 30, 22.5, 18, and 15 degrees between them, supporting more than seven representative classes of face direction. Said SIT’s Chinthaka Premachandra, “Precise training data for each orientation was obtained from the integration of the depth and gyro sensors, which reduce the number of point cloud samples required for constructing the classification model. Furthermore, applying a weight reduction process to reduce the weight of point cloud data enhanced training efficiency and resulted in fast face orientation estimation.”

From “Face Orientation Estimation Using Depth-Gyro Sensor”
Shibaura Institute of Technology (Japan) (09/22/23)
View Full Article