# US 2014/0195472 A1 · Information processing apparatus, generating method, medical diagnosis support apparatus, and medical diagnosis support method > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2014/0195472 A1 - Document ID: US20140195472A1 - Status: Published application - Kind: A1 - Application number: 14/147,898 - Filed: 2014-01-06 - Published: 2014-07-10 - Earliest claimed: 2013-01-09 - Assignees: Canon Kabushiki Kaisha - Applicants: Canon Kabushiki Kaisha - Inventors: Masami Kawagishi - Web page: https://patentgrep.com/patent/US20140195472A1 ## CPC - G06N 5/04: Inference or reasoning models - G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems - G16Z 99/00: Subject matter not provided for in other main groups of this subclass - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA - G16Z: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR ## Abstract A medical diagnosis support apparatus includes a training data obtaining unit that obtains training data, an inference means candidate creating unit that creates a plurality of inference means candidates based on the training data, an inference performance evaluation unit that evaluates the performance of the plurality of inference means candidates based on the training data, an information validity evaluation unit that evaluates the validity of information presented by each of the plurality of inference means candidates based on the training data, and an inference means selection unit that selects an inference means from the plurality of inference means candidates based on the performance of the plurality of inference means candidates and the validity of the information presented by each of the plurality of inference means candidates. ## Claims ### Claim 1 An information processing apparatus providing an inference means for outputting a diagnosis inferred for a medical case and reason information used to infer the diagnosis, the apparatus comprising: a first obtainment unit configured to obtain values indicating an inference performance for each of a plurality of inference means based on a group of first correct data including a correct diagnosis for the medical case; a second obtainment unit configured to obtain values indicating a validity of the reason information used in the inference output by each of the plurality of inference means, based on a group of second correct data that includes information to be identified in a case for which the correct diagnosis has been provided; and a third obtainment unit configured to obtain at least one inference means from the plurality of inference means based on the values indicating the inference performance and the values indicating the validity. ### Claim 2 (depends on 1) The information processing apparatus according to claim 1, further comprising: a display control unit configured to display, in a display unit, the diagnosis inferred by the inference means obtained by the third obtainment unit and the reason information used to infer that diagnosis based on a plurality of pieces of identified information identified for a given case. ### Claim 3 An information processing apparatus providing an inference means for, based on a plurality of pieces of information identified for a medical case, outputting a diagnosis for the case, and outputting at least one of the identified pieces of information as information having the most influence on the inference of the diagnosis, the apparatus comprising: an obtainment unit configured to obtain a group of correct data including a correct diagnosis for the case and information to be identified for the case having that correct diagnosis; a calculation unit configured to obtain a value indicating the validity of information output as information having the most influence on the inference made by each of a plurality of inference means based on the group of correct data; and a selection unit configured to select at least one inference means from the plurality of inference means based on the value indicating inference performance and the value indicating the validity. ### Claim 4 An information processing apparatus providing an inference means that has structured data in which a plurality of elements and a diagnosis are numerically associated with each other and that can infer a diagnosis for a given medical case by determining which of the plurality of elements a plurality of pieces of information identified for the case apply to, the apparatus comprising: a first obtainment unit configured to obtain values indicating the inference performance of each of a plurality of inference means that infer a diagnosis for the medical case, based on a group of first correct data in which the identified information and a correct diagnosis are associated with each other; a second obtainment unit configured to obtain values indicating the validity of at least one of the elements used in the inference performed by the plurality of inference means, based on a group of second correct data in which a correct diagnosis is associated with a plurality of pieces of information to be identified for a case having that correct diagnosis; and a selection unit configured to select at least one inference means from the plurality of inference means based on the values indicating the inference performance and the values indicating the validity. ### Claim 5 (depends on 4) The information processing apparatus according to claim 4, further comprising: a display control unit configured to display, based on a plurality of pieces of information identified for a given case, a diagnosis for the case obtained through an inference made by the inference means selected by the selection unit and the information among the plurality of pieces of identified information that has the most influence on the inference of that diagnosis, in a display unit. ### Claim 6 A generating method for generating an inference means that outputs a diagnosis inferred for a medical case and reason information used to infer the diagnosis, the method comprising: a first obtainment step of obtaining values indicating an inference performance for each of a plurality of inference means based on a group of first correct data including a correct diagnosis for the medical case; a second obtainment step of obtaining values indicating a validity of the reason information used in the inference output by each of the plurality of inference means, based on a group of second correct data that includes information to be identified in a case for which the correct diagnosis has been provided; and a generating step of generating at least one inference means from the plurality of inference means based on the values indicating the inference performance and the values indicating the validity. ### Claim 7 A generating method for generating an inference means that, based on a plurality of pieces of information identified for a medical case, outputs a diagnosis for the case, and outputs at least one of the identified pieces of information as information having the most influence on the inference of the diagnosis, the method comprising: an obtainment step of obtaining a group of correct data including a correct diagnosis for the case and information to be identified for the case having that correct diagnosis; a calculation step of obtaining a value indicating the validity of information output as information having the most influence on the inference made by each of a plurality of inference means based on the group of correct data; and a generating step of generating at least one inference means from the plurality of inference means based on the value indicating the inference performance and the value indicating the validity. ### Claim 8 A generating method for generating an inference means that has structured data in which a plurality of elements and a diagnosis are numerically associated with each other and that can infer a diagnosis for a given medical case by determining which of the plurality of elements a plurality of pieces of information identified for the case apply to, the apparatus comprising: a first obtainment step of obtaining values indicating the inference performance of each of a plurality of inference means that infer a diagnosis for the medical case, based on a group of first correct data in which the identified information and a correct diagnosis are associated with each other; a second obtainment step of obtaining values indicating the validity of at least one of the elements used in the inference performed by the plurality of inference means, based on a group of second correct data in which a correct diagnosis is associated with a plurality of pieces of information to be identified for a case having that correct diagnosis; and a generating step of generating at least one inference means from the plurality of inference means based on the values indicating the inference performance and the values indicating the validity. ### Claim 9 A medical diagnosis support apparatus comprising: a training data obtainment unit configured to obtain training data; a candidate creating unit configured to create a plurality of inference means candidates based on the training data; an inference performance evaluation unit configured to evaluate the performance of the plurality of inference means candidates based on the training data; an information validity evaluation unit configured to evaluate the validity of information presented by each of the plurality of inference means candidates based on the training data; and a selection unit configured to select an inference means from the plurality of inference means candidates based on the performance of the plurality of inference means candidates and the validity of the information presented by each of the plurality of inference means candidates. ### Claim 10 (depends on 9) The medical diagnosis support apparatus according to claim 9, wherein the candidate creating unit is operable to create the inference means candidate as various inference methods. ### Claim 11 (depends on 9) The medical diagnosis support apparatus according to claim 9, wherein the candidate creating unit is operable to create the inference means candidates as Bayesian networks having different graph structures. ### Claim 12 (depends on 9) The medical diagnosis support apparatus according to claim 9, wherein the candidate creating unit is operable to create the plurality of inference means candidates through processing using the Markov chain Monte Carlo method; and the selection unit is operable to select the inference means from the plurality of inference means candidates through the processing. ### Claim 13 (depends on 9) The medical diagnosis support apparatus according to claim 9, wherein the candidate creating unit is operable to create the plurality of inference means candidates through processing using a genetic algorithm; and the selection unit is operable to select the inference means from the plurality of inference means candidates through the processing. ### Claim 14 (depends on 9) The medical diagnosis support apparatus according to claim 9, further comprising: an additional data obtainment unit configured to obtain additional data, wherein the inference performance evaluation unit is operable to evaluate the performance of the plurality of inference means candidates based on the training data and the additional data; and the information validity evaluation unit is operable to evaluate the information presented by the plurality of inference means candidates based on the training data and the additional data. ### Claim 15 (depends on 14) The medical diagnosis support apparatus according to claim 14, wherein the additional data obtainment unit is operable to obtain additional data while the medical diagnosis support apparatus is operating. ### Claim 16 (depends on 14) The medical diagnosis support apparatus according to claim 14, wherein the additional data obtainment unit is operable to obtain the additional data at a predetermined timing. ### Claim 17 (depends on 14) The medical diagnosis support apparatus according to claim 14, wherein the additional data obtainment unit is operable to obtain the additional data from a database different from a database from which the training data is obtained. ### Claim 18 (depends on 9) The medical diagnosis support apparatus according to claim 9, wherein the information validity evaluation unit is operable to evaluate the validity of the information by changing processing based on the performance of the inference means candidates. ### Claim 19 (depends on 9) The medical diagnosis support apparatus according to claim 9, wherein the information validity evaluation unit is configured not to evaluate the validity of the information in the case where the performance of the inference means candidates exceeds a threshold. ### Claim 20 (depends on 11) The medical diagnosis support apparatus according to claim 11, wherein the candidate creating unit is operable to modify the graph structure based on distances from a diagnosis node in the graph structure. ### Claim 21 (depends on 11) The medical diagnosis support apparatus according to claim 11, further comprising: a presenting unit configured to present information supporting a medical diagnosis based on partial information of medical information created in accordance with the graph structure. ### Claim 22 (depends on 21) The medical diagnosis support apparatus according to claim 21, further comprising: a creating unit configured to create the partial information based on distances from a diagnosis node in the graph structure. ### Claim 23 (depends on 22) The medical diagnosis support apparatus according to claim 22, wherein the creating unit is operable to create the partial information based on distances between nodes in the graph structure. ### Claim 24 (depends on 9) The medical diagnosis support apparatus according to claim 9, wherein the candidate creating unit is operable to set weights for a plurality of inference means candidates created using different parameters and to construct an inference means by integrating the plurality of inference means candidates based on the weights. ### Claim 25 A medical diagnosis support method for a medical diagnosis support apparatus, the method comprising: an obtainment step of obtaining training data; a creating step of creating a plurality of inference means candidates based on the training data; an inference performance evaluation step of evaluating the performance of the plurality of inference means candidates based on the training data; an information validity evaluation step of evaluating the validity of information presented by each of the plurality of inference means candidates based on the training data; and a selection step of selecting an inference means from the plurality of inference means candidates based on the performance of the plurality of inference means candidates and the validity of the information presented by each of the plurality of inference means candidates.