US 12,249,124 B2
Super resolution device and method
Kung Ho Lee, Hsinchu (TW); Yu Cheng Cheng, Hsinchu (TW); and Hsu-Tung Shih, Hsinchu (TW)
Assigned to Realtek Semiconductor Corporation, Hsinchu (TW)
Filed by Realtek Semiconductor Corporation, Hsinchu (TW)
Filed on Sep. 30, 2022, as Appl. No. 17/936,851.
Claims priority of application No. 111117909 (TW), filed on May 12, 2022.
Prior Publication US 2023/0368496 A1, Nov. 16, 2023
Int. Cl. G06T 17/00 (2006.01); G06T 3/4053 (2024.01); G06V 10/77 (2022.01)
CPC G06V 10/7715 (2022.01) [G06T 3/4053 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A super resolution device, comprising:
an input interface, being configured to receive an image and a scaling signal, wherein the scaling signal is configured to indicate to perform a double scaling operation or a quadruple scaling operation to the image; and
a sharing layer calculator group, comprising a plurality of multiply-accumulate calculators corresponding to a first number, wherein the multiply-accumulate calculators are configured to perform a plurality of convolution operations, and the first number is twice as big as an initial number of output channels, and each of the multiply-accumulate calculators performs a plurality of corresponding multiplication operations and addition operations through a fixed number of hardware components;
wherein the sharing layer calculator group performs the double scaling operation or the quadruple scaling operation based on the scaling signal, and the sharing layer calculator group jointly uses the same number of the multiply-accumulate calculators to perform the double scaling operation and the quadruple scaling operation to the image;
wherein the sharing layer calculator group performs the convolution operations based on a first number of input channels and a first number of output channels when the scaling signal indicates to perform the quadruple scaling operation, the first number of input channels and the first number of output channels are at least an even positive integer greater than or equal to 2;
wherein the sharing layer calculator group performs the convolution operations based on a second number of input channels and a second number of output channels when the scaling signal indicates to perform the double scaling operation;
wherein the first number of input channels is twice as big as the second number of input channels, and the first number of output channels is twice as big as the second number of output channels.