US 11,055,240 B2
Data processing method and device for performing a convolution operation
Bo Wen, Guangdong (CN); Qingxin Cao, Guangdong (CN); and Wei Li, Guangdong (CN)
Assigned to Shenzhen Intellifusion Technologies Co., Ltd., Shenzhen (CN)
Appl. No. 17/253,625
Filed by Shenzhen Intellifusion Technologies Co., Ltd., Shenzhen (CN)
PCT Filed Oct. 30, 2019, PCT No. PCT/CN2019/114332
§ 371(c)(1), (2) Date Dec. 18, 2020,
PCT Pub. No. WO2020/134529, PCT Pub. Date Jul. 2, 2020.
Claims priority of application No. 201811654890.2 (CN), filed on Dec. 29, 2018.
Prior Publication US 2021/0124698 A1, Apr. 29, 2021
Int. Cl. G06F 13/16 (2006.01); G06T 1/60 (2006.01); G06F 5/12 (2006.01); G06F 17/15 (2006.01)
CPC G06F 13/1673 (2013.01) [G06F 5/12 (2013.01); G06F 13/1678 (2013.01); G06F 17/15 (2013.01); G06T 1/60 (2013.01)] 9 Claims
OG exemplary drawing
 
1. A data processing method comprising:
if it is detected that a number of image data to be transferred is greater than zero wherein the number of image data is a product of a number of input image data and a number of output image data, and a first available storage space of a first first-in first-out (FIFO) memory is greater than or equal to a storage space occupied by an N number of input image data, transferring the N input image data in a first memory to the first FIFO memory;
if it is detected that a number of weight data to be transferred wherein the number of weight data is a product of the number of input image data and the number of output image data, is greater than zero and a second available storage space of a second FIFO memory is greater than or equal to a storage space occupied by an M number of weight data, transferring the M weight data in a second memory to the second FIFO memory; and wherein each weight data comprises input indexes of input image data and output indexes of output image data; and
when the number of input image data cached in the first FIFO memory and the number of weight data cached in the second FIFO memory are respectively greater than or equal to 1, performing a convolution operation on input image data i read from the first FIFO memory and weight data w read from the second FIFO memory, to obtain output image data corresponding to an output index of the weight data w, and an input index of the weight data w corresponding to the input image data i; and wherein N, M, i, and w are integers.