US 12,387,826 B2
Document creation apparatus, document creation method, and document creation program for creating document based on medical image
Keigo Nakamura, Tokyo (JP); and Yuya Hamaguchi, Tokyo (JP)
Assigned to FUJIFILM Corporation, Tokyo (JP)
Filed by FUJIFILM Corporation, Tokyo (JP)
Filed on Sep. 3, 2023, as Appl. No. 18/460,583.
Application 18/460,583 is a continuation of application No. PCT/JP2022/002462, filed on Jan. 24, 2022.
Claims priority of application No. 2021-046614 (JP), filed on Mar. 19, 2021; and application No. 2021-208719 (JP), filed on Dec. 22, 2021.
Prior Publication US 2023/0420096 A1, Dec. 28, 2023
Int. Cl. G06F 40/186 (2020.01); G06F 40/40 (2020.01); G06T 7/00 (2017.01); G16H 15/00 (2018.01)
CPC G16H 15/00 (2018.01) [G06F 40/186 (2020.01); G06F 40/40 (2020.01); G06T 7/0012 (2013.01)] 9 Claims
OG exemplary drawing
 
1. A document creation apparatus comprising:
a memory that stores a plurality of creation methods comprising a plurality of templates and a plurality of trained models corresponding to different natural language processing schemes; and
at least one processor configured to:
perform image analysis on a medical image to acquire an analysis result including one or more findings;
determine a number of the findings included in the analysis result so as to determine a creation method of a document to be created, wherein the number of the findings is associated with templates or natural language processing schemes;
in response to the number of the findings being less than a predetermined threshold, determine that the creation method of the document to be created being one template corresponding to the number of findings among the plurality of templates, and create a finding remark by inserting the findings into the one template;
in response to the number of findings being equal to or greater than the predetermined threshold, determine that the creation method of the document to be created being one natural language processing scheme corresponding to the number of findings among the plurality of natural language processing schemes, and create the finding remark by inputting the findings into the trained model among the plurality of trained models corresponding to the one natural language processing scheme; and
create the document by using the created finding remark including a graphical display of the analysis result.