A configuration of the image estimation processing unit 902 in the present embodiment will be described below. Unlike in embodiment 1, the image estimation processing unit 902 includes a single neural network. FIG. 10 shows an example configuration of the neural network in the image estimation processing unit 902. Although the process in the present embodiment will be described below based on the neural network shown in FIG. 10, the technique of the present disclosure is not limited to it.

First, a convolution layer 1001 performs convolution operation on 256×256×4 input image data. For the pixel value of the input image data at a position (x, y) in a p plane being I(x, y, p1), output image data I_G1(x, y, p2) from the convolution layer 1001 is calculated by the equation below.

$\begin{array}{c}{I}_{G??1}?\left(x,y\right)=\sum _{p=1}^{p=4}?\underset{}{\overset{}{}}\end{array}$