Fig. 4. Building an optical CNN for imaging recognition.

a Operation procedure of using the OCNN to recognize handwriting numbers from the MNIST database. The OCNN consists of a convolution layer with two kernels, a pooling and a fully connected layer. The output gives the answer whether the input image is “1” or “2”. b The convolution kernel matrices K₁ and K₂ generated by training the OCNN. c. Raw output data of the convolution layer of two kernel matrices. d The weight bank matrix used in the fully connected layer. e The recognition results from the experiment with the OCNN (left) and the calculation with a computer (right) show an excellent agreement.