HP INCFollow


This disclosure describes a realtime 2-D image-based human face-to-camera distance estimation and facial dimension measurement system, which is without the requirement of user calibration. The proposed system first uses 2 deep-learning networks for human face localization and facial landmarks detection. Then, the proposed system utilizes facial landmarks generated from the first step to compute the distance between the human face and the camera. The proposed system can be easily extended to generate facial dimensions in realtime. The camera is pre-calibrated. Reference human facial landmarks and their corresponding scale, as well as the camera calibration information, are provided as prior knowledge. The proposed system is based on only regular 2-D images without thermal, stereo, and depth sensor information.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Share Alike 4.0 License.