IP Library › Granted Patent US 10,930,396
Granted Patent B2
US 10,930,396 · App. 16/440,830 · Granted Feb 23, 2021

Similar case search apparatus, method for operating similar case search apparatus, and similar case search system

Inventor: Shoji Kanada (Tokyo, JP)
Assignee: FUJIFILM Corporation
G16H50/20A61B6/03G06F16/55G06Q50/22G06T7/0014G16H10/00G16H30/00G16H30/40G16H50/70
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 10,930,396
App. No.
16/440,830
Granted
Feb 23, 2021
Kind
B2
Abstract

A weighting processing unit performs a weighting process for an individual lesion similarity which is a similarity between a target lesion and a case lesion of the same type, using a first weighting coefficient corresponding to the degree of contribution of each type of lesion to the specification of a disease name, to calculate a processed individual lesion similarity corresponding to a first purpose of specifying the disease name. In addition, the weighting processing unit performs the weighting process for the individual lesion similarity, using a second weighting coefficient corresponding to the degree of contribution of each type of lesion to the specification of a severity, to calculate a processed individual lesion similarity corresponding to a second purpose of specifying the severity. A total similarity between a target image and a case image is calculated for each purpose on the basis of the processed individual lesion similarities.

Claims (61)

1. A similar case search apparatus that searches for similar cases including a case image similar to a target image which is a medical examination target from a plurality of cases including the case image, comprising:

a processor configured to:

acquire a type of target lesion which is a lesion present in the target image;

acquire a type of case lesion which is a lesion present in the case image;

acquire a feature amount of the target lesion;

acquire a feature amount of the case lesion;

calculate an individual lesion similarity which is a similarity between the target lesion and the case lesion of the same type on the basis of the feature amount of the target lesion and the feature amount of the case lesion;

perform a weighting process for the individual lesion similarity, using weighting coefficients which are preset for each of the types and a plurality of purposes;

calculate a total similarity between the target image and the case image for each of the purposes on the basis of the individual lesion similarity subjected to the weighting process; and

search for a similar case on the basis of the total similarity for at least one of the purposes,

wherein the purposes include a purpose of specifying a disease name and a purpose of specifying a severity of a disease, and

the processor is further configured to:

perform the weighting process, using a first weighting coefficient which is the weighting coefficient corresponding to a degree of contribution of each type to the specification of the disease name, and

perform the weighting process, using a second weighting coefficient which is the weighting coefficient corresponding to a degree of contribution of each type to the specification of the severity.

2. The similar case search apparatus according to claim 1 ,

wherein the weighting coefficients for the types and the purposes are set for each disease.

3. The similar case search apparatus according to claim 1 ,

wherein the case includes the disease name, and

the first weighting coefficient is set on the basis of a first likelihood ratio obtained by statistically analyzing a causal relationship between the case lesion and the disease name included in the case.

4. The similar case search apparatus according to claim 1 ,

wherein the case includes the severity, and

the second weighting coefficient is set on the basis of a second likelihood ratio obtained by statistically analyzing a causal relationship between the case lesion and the severity included in the case.

5. The similar case search apparatus according to claim 1 ,

wherein the processor is further configured to create a list of the searched similar cases for each of the purposes.

6. The similar case search apparatus according to claim 5 ,

wherein the processor is further configured to:

output a list display screen for displaying the list,

display a first list which is the list corresponding to a first purpose and from which one of the similar cases is selected, and

display a second list which is the list corresponding to a second purpose in a case in which one of the similar cases in the first list is selected on the list display screen.

7. The similar case search apparatus according to claim 6 ,

wherein the first purpose is to specify a disease name and the second purpose is to specify the severity of a disease.

8. The similar case search apparatus according to claim 7 ,

wherein the similar cases are arranged in the first list according to a candidate disease ranking which is a ranking of the disease names included in the similar cases on the basis of the total similarity corresponding to the first purpose, and

the similar cases are arranged in the second list according to a ranking in the same disease which is a ranking of the total similarity corresponding to the second purpose in the similar cases of the same disease name.

9. A method for operating a similar case search apparatus that searches for similar cases including a case image similar to a target image which is a medical examination target from a plurality of cases including the case image, the method comprising:

a first type acquisition step of acquiring a type of target lesion which is a lesion present in the target image;

a second type acquisition step of acquiring a type of case lesion which is a lesion present in the case image;

a first feature amount acquisition step of acquiring a feature amount of the target lesion;

a second feature amount acquisition step of acquiring a feature amount of the case lesion;

an individual lesion similarity calculation step of calculating an individual lesion similarity which is a similarity between the target lesion and the case lesion of the same type on the basis of the feature amount of the target lesion and the feature amount of the case lesion;

a weighting processing step of performing a weighting process for the individual lesion similarity, using weighting coefficients which are preset for each of the types and a plurality of purposes;

a total similarity calculation step of calculating a total similarity between the target image and the case image for each of the purposes on the basis of the individual lesion similarity subjected to the weighting process; and

a search step of searching for a similar case on the basis of the total similarity for at least one of the purposes,

wherein the purposes include a purpose of specifying a disease name and a purpose of specifying a severity of a disease, and

the method further comprises:

performing the weighting process, using a first weighting coefficient which is the weighting coefficient corresponding to a degree of contribution of each type to the specification of the disease name, and

performing the weighting process, using a second weighting coefficient which is the weighting coefficient corresponding to a degree of contribution of each type to the specification of the severity.

10. A similar case search system including a similar case search apparatus that searches for similar cases including a case image similar to a target image which is a medical examination target from a plurality of cases including the case image and a client terminal that is connected to the similar case search apparatus by a network and is operated by a doctor, the system comprising:

a processor is configured to:

acquire a type of target lesion which is a lesion present in the target image;

acquire a type of case lesion which is a lesion present in the case image;

acquire a feature amount of the target lesion;

acquire a feature amount of the case lesion;

calculate an individual lesion similarity which is a similarity between the target lesion and the case lesion of the same type on the basis of the feature amount of the target lesion and the feature amount of the case lesion;

perform a weighting process for the individual lesion similarity, using weighting coefficients which are preset for each of the types and a plurality of purposes;

calculate a total similarity between the target image and the case image for each of the purposes on the basis of the individual lesion similarity subjected to the weighting process; and

search for a similar case on the basis of the total similarity for at least one of the purposes,

wherein the purposes include a purpose of specifying a disease name and a purpose of specifying a severity of a disease, and

the processor is further configured to:

perform the weighting process, using a first weighting coefficient which is the weighting coefficient corresponding to a degree of contribution of each type to the specification of the disease name, and

perform the weighting process, using a second weighting coefficient which is the weighting coefficient corresponding to a degree of contribution of each type to the specification of the severity.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 13, 2019
From: KANADA, SHOJI
To: FUJIFILM CORPORATION
Reel/Frame 049464/0621 →
Priority Claims (1)
JP JP2016-245304 · Dec 19, 2016 · national
Continuity (2)
Continuation PCTJP2017041768 · Nov 21, 2017
Related Publication 20190295724A1 · Sep 26, 2019
Cited By (2)
US 12,299,888 US 12,541,967