US 11,937,971 B2
Method for classifying disease using artificial intelligence and electronic apparatus therefor
Jung Ho Lee, Seoul (KR); Won Yang Cho, Gyeonggi-do (KR); and Eun Joo Lee, Seoul (KR)
Assigned to SMARTSOUND CORPORATION, Seoul (KR)
Filed by SMARTSOUND CORPORATION, Seoul (KR)
Filed on May 5, 2023, as Appl. No. 18/143,919.
Claims priority of application No. 10-2022-0090887 (KR), filed on Jul. 22, 2022.
Prior Publication US 2024/0023922 A1, Jan. 25, 2024
Int. Cl. A61B 7/00 (2006.01); A61B 5/00 (2006.01); G16H 10/60 (2018.01)
CPC A61B 7/003 (2013.01) [A61B 5/7267 (2013.01); G16H 10/60 (2018.01)] 18 Claims
OG exemplary drawing
 
1. A method, performed by an electronic apparatus, for classifying heart and lung related diseases using artificial intelligence (AI), the method comprising:
obtaining auscultation sound data, auscultation position data and setting data including information regarding body part;
obtaining feature information based on the auscultation sound data and obtaining auscultation position information based on the auscultation position data;
generating combined information by combining the feature information and the auscultation position information; and
identifying at least one of heart and lung related disease information corresponding to the combined information, by inputting the combined information to an AI model,
wherein the feature information, the auscultation position information and the combined information have a vector form,
wherein the obtaining of the auscultation position information comprises obtaining the auscultation position information by embedding the auscultation position data,
wherein the generating of the combined information comprises generating the combined information by concatenating the feature information and the auscultation position information,
wherein the auscultation position information includes configured number of components corresponding to possible auscultation positions,
wherein, from among the configured number of components corresponding to the possible auscultation positions, a component corresponding to the auscultation position has a first value and other remaining components have a second value, and
wherein the configured number is determined based on the information regarding body part.