A Dual Discriminator Adversarial Learning Approach for Dental Occlusal Surface Reconstruction
Algorithm 2
Depth map calculation.
Input: h is the center point of the bounding box to a projection plane.
Output: pixel is the pixel value of the corresponding point on the tooth model.
Step 1: a projection plane is constructed at the vertical distance h of the bounding box with the center of the tooth model as the origin.
Step 2: the projection plane is divided into 256 × 256 small grids, and the corresponding relationship with the pixel position (i, j) of the depth map is established.
Step 3: the shortest distance d from each grid center point P(i, j) on the projection plane to the occlusal surface is calculated by the ray intersection method.
Step 4:
the calculated distance d is converted to the range of 0–255 pixel values.