Hello. If NN uses for 3d objects, the 3D matrix “cube of the numbers” will be transformed to the vector? Which algorithm for the 3D matrix to the vector will be used?
Usually what is done for Forward Neural Networks is you Flatten the data before you input it into the network so you will get 1D Data.
Further, there will neural layer called Convolution Layer which can be 1D, 2D, or 3D. the 3D is usually applied to 3D CT Scans like detecting brain cancer in CT Scans. this is one of the applications I saw before applied at 3D data.