Implicit Camera Calibration Using MultiLayer Perceptron Type Neural Network

This paper suggest a new camera calibration approach based on the neural network model. The proposed approach is shown to be very accurate because the neural network model implicitly contains all the physical parameters, some of which are very difficult to be estimated in the conventional explicit calibration methods. As the first step of this approach, this paper presents the camera calibration process which enables the coordinate transformation between 2D image points and points of a certain space in 3D real world. However, this approach is currently extended to be a general 3D camera calibration method in terms of 2 plane method. Experimental comparison of our method with well-known Tsai's 2 stage method is made to verify the accuracy of the proposed method.