Bayesian modeling of 3D shape inference from line drawings

Human depth comparisons in line drawings reflect the underlying uncertainty of perceived 3D shape. propose a Bayesian model that estimates the 3D shape from line drawings based on the local and local contour cues. This model estimates the posterior distribution over depth differences at two points a line drawing. The likelihood is numerically co which generates random 3D surfaces and, via projection, random line drawings. The 3D surfaces inflated from random skeletons and projected into line drawings. Given a novel line drawing, the samples probable local surfaces based on the relations between local 3D surface patches corresponding 2D contour segments. Then, the likelihood function of depth differences is estimated the distribution of probable surface orientations (Figure 2). The known human biases in depth perception, such as slant figure/ground organization. This model predicts the probabilities assigned to depth differences two points on line drawings from the posterior on depth differences (Figure 3). These probabilities consistent with human responses (Figure 4), showing that the model accounts for human interpretation line drawings. This model encodes the uncertainty in 3D shape interpretation from line simulates the propagation of depth information from local and global contours, and provides a tool testing the scope of cues in 3D shape inference (Figure