Human-like monocular depth biases in deep neural networks

Fig 6

Exponential fitting coefficients for 31 DNNs that produce absolute depth values.

The figure consists of three subplots: scale component (C), exponent component (γ), and shift component (β).