Neural network for enhancing original image, and computer-implemented method for enhancing original image using neural network

A computer-implemented method for enhancing an original image using a neural network, comprising:

• inputting an original image to a neural network having 2n number of sampling units sequentially connected, and a plurality of processing units, wherein a respective one of the plurality of processing units is between two adjacent sampling units of the 2n number of sampling units;
• and
• processing the original image using the 2n number of sampling units, and the plurality of processing units, thereby obtaining an enhanced image;
• wherein a first sampling unit to an n-th sample unit of the 2n number of sampling units are DeMux units;
• an (n+1)-th sample unit to a (2n)-th sample unit of the 2n number of sampling units are Mux units;
• and
• the respective one of the plurality of processing units comprises a plurality of convolutional blocks sequentially connected;
• wherein, in a respective one of the DeMux units, the computer-implemented method comprises rearranging pixels in a respective input image to the respective one of the DeMux units following a first scrambling rule to obtain a respective rearranged image;
• and dividing the respective rearranged image into m number of divided images;
• wherein, in a respective one of the Mux units, the computer-implemented method comprises combining respective m′ number of input images to the respective one of the Mux units to obtain a respective combined image;
• and rearranging pixels of the respective combined image following a second scramble rule to obtain an output image;
• wherein m is an integer equivalent to or greater than 2, m′ is an integer equivalent to or greater than 2, and n is a positive integer.

A computer-program product comprising non-transitory tangible computer-readable medium having computer-readable instructions thereon, the computer-readable instructions being executable by a processor to cause the processor to perform:

• inputting an original image to a neural network having 2n number of sampling units sequentially connected, and a plurality of processing units, wherein a respective one of the plurality of processing units is between two adjacent sampling units of the 2n number of sampling units;
• and
• processing the original image using the 2n number of sampling units, and the plurality of processing units, thereby obtaining an enhanced image;
• wherein a first sampling unit to an n-th sample unit of the 2n number of sampling units are DeMux units;
• an (n+1)-th sample unit to a (2n)-th sample unit of the 2n number of sampling units are Mux units;
• and
• the respective one of the plurality of processing units comprises a plurality of convolutional blocks sequentially connected;
• wherein, in a respective one of the DeMux units, the computer-implemented method comprises rearranging pixels in a respective input image to the respective one of the DeMux units following a first scrambling rule to obtain a respective rearranged image;
• and dividing the respective rearranged image into m number of divided images;
• wherein, in a respective one of the Mux units, the computer-implemented method comprises combining respective m′ number of input images to the respective one of the Mux units to obtain a respective combined image;
• and rearranging pixels of the respective combined image following a second scramble rule to obtain an output image;
• wherein m is an integer equivalent to or greater than 2, m′ is an integer equivalent to or greater than 2, and n is a positive integer.