There are three main applications of ToF cameras in the car. One is the face recognition or Face ID on the car’s B-pillar, the second is the gesture recognition Gesture of the car’s central control, and the third is the driver status monitoring DMS. In the future, occupant detection may be added to the presence of people in the car.
2D flat camera cannot do Face ID and gesture control. From the mobile phone field, it can be seen that no one dares to do this. Without depth data, a 2D flat camera can be easily deceived. There is no need to say more about gesture control. In the field of DMS, 2D plane cameras also face many challenges. One is that strong light or rapid light changes, such as the sun shining on the camera, boulevards and other scenes, will blind or fail to respond to the 2D plane camera. The second is the algorithm. Deep learning models are getting bigger and bigger, and more and more algorithm resources are consumed, which means that the cost of hardware processors is getting higher and higher. The third is accuracy. There is no deep data, or the deep data inferred by deep learning not only has low accuracy and consumes a lot of computing resources. A 2D flat camera can usually only count the number of blinks and the degree of eye-lid opening and closing.
3D face recognition processes 3D data, such as point clouds, voxels, etc. These data are complete, three-dimensional, highly accurate, and do not require deep learning convolution, the algorithm is simple and reliable, and the robustness is high.so ToF cameras is the perfect alternative to a 2D camera.
