US 12,243,320 B2
Information processing method, program, and information processing apparatus
Haruyoshi Yonekawa, Kanagawa (JP)
Assigned to Sony Semiconductor Solutions Corporation, Kanagawa (JP)
Appl. No. 17/442,075
Filed by Sony Semiconductor Solutions Corporation, Kanagawa (JP)
PCT Filed Mar. 17, 2020, PCT No. PCT/JP2020/011601
§ 371(c)(1), (2) Date Sep. 22, 2021,
PCT Pub. No. WO2020/203241, PCT Pub. Date Oct. 8, 2020.
Claims priority of application No. 2019-065378 (JP), filed on Mar. 29, 2019.
Prior Publication US 2022/0172484 A1, Jun. 2, 2022
Int. Cl. G06V 20/56 (2022.01); G01S 13/89 (2006.01); G06N 3/08 (2023.01); G06V 10/82 (2022.01)
CPC G06V 20/56 (2022.01) [G01S 13/89 (2013.01); G06N 3/08 (2013.01); G06V 10/82 (2022.01)] 14 Claims
OG exemplary drawing
 
1. An information processing method executed by processing circuitry, the method comprising:
generating feature data that numerically represents a feature of a feature map generated from input data in a model using a neural network;
generating analysis data based on the feature data of a plurality of feature maps; and
displaying the generated analysis data as feature data, based on feature maps generated from a same input data, arranged in numerical order of the feature maps in a first axis direction and feature data, based on feature maps corresponding to different input data, arranged in numerical order of the different input data in a second axis direction orthogonal to the first axis direction, wherein the generated analysis data is displayed as an image in which pixels represent values of the feature data of a given input data and a given feature map, and the image is indicative of a state of learning of the model using the neural network.