IP Library Granted Patent US 12705963
Granted Patent B2
US 12705963 · App. 18/809,478 · Granted Aug 11, 2026

Learning apparatus, estimation apparatus, learning method, and non-transitory storage medium

Inventors: Jianquan Liu (Tokyo, JP); Kenta Ishihara (Tokyo, JP)
Assignee: NEC CORPORATION
G08B13/19613G08B13/19604
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 12705963
App. No.
18/809,478
Granted
Aug 11, 2026
Kind
B2
Abstract

The present invention provides a learning apparatus ( 10 ) including: an acquisition unit ( 11 ) that acquires an image; a similarity computation unit ( 12 ) that computes a similarity between the acquired image, and a first image being accumulated in advance and indicating an abnormal state; a registration unit ( 13 ) that registers, as a second image indicating a normal state, the acquired image whose similarity is equal to or less than a first reference value; and a learning unit ( 14 ) that generates an estimation model for discriminating between normal and abnormal by machine learning using the first image and the second image.

Claims (42)

1 . A learning apparatus comprising:

at least one memory storing instructions; and

at least one processor configured to execute the instructions to:

generate a first learning model for discriminating between a first category and a second category by performing supervised learning based on a first image categorized as the first category and a second image categorized as the second category;

use the first learning model to acquire a category of a first captured image as one of a plurality of categories, the plurality of categories including the first category and the second category;

receive a determination indicating whether the acquired category is either correct or incorrect;

generate a second learning model by performing supervised learning based on the first captured image and the determination, wherein the second learning model categorizes an image into the plurality of categories; and

use the second learning model to acquire a category of a second captured image as one of the plurality of categories.

2 . The learning apparatus according to claim 1 , wherein the at least one processor configured to execute the instructions to control a display apparatus to display the first captured image that is categorized as the first category by the first learning model.

3 . The learning apparatus according to claim 2 , wherein the displayed first captured image is categorized as the first category with reliability equal to or higher than a predetermined level in the first learning model.

4 . The learning apparatus according to claim 1 , wherein the determination is based on an input by a user via a display apparatus that displays the first captured image.

5 . The learning apparatus according to claim 1 , wherein the first image is accumulated previously, wherein the second image is determined as the first category by comparing with the first image.

6 . The learning apparatus according to claim 1 , wherein the first category is anomaly and the second category is normal.

7 . The learning apparatus according to claim 6 ,

wherein the at least one processor configured to execute the instructions to generate the first learning model by performing supervised learning based on a plurality of first images categorized as the first category and a plurality of second images categorized as the second category, and

wherein the plurality of first images comprises images identified by a user as anomaly.

8 . A learning method executed by a computer, the learning method comprising:

generating a first learning model for discriminating between a first category and a second category by performing supervised learning based on a first image categorized as the first category and a second image categorized as the second category;

using the first learning model to acquire a category of a first captured image as one of a plurality of categories, the plurality of categories including the first category and the second category;

receiving a determination indicating whether the acquired category is either correct or incorrect;

generating a second learning model by performing supervised learning based on the first captured image and the determination, wherein the second learning model categorizes an image into the plurality of categories; and

using the second learning model to acquire a category of a second captured image as one of the plurality of categories.

9 . The learning method according to claim 8 , further comprising controlling a display apparatus to display the first captured image that is categorized as the first category by the first learning model.

10 . The learning method according to claim 9 , wherein the displayed first captured image is categorized as the first category with reliability equal to or higher than a predetermined level in the first learning model.

11 . The learning method according to claim 8 , wherein the determination is based on an input by a user via a display apparatus that displays the first captured image.

12 . The learning method according to claim 8 , wherein the first category is anomaly and the second category is normal.

13 . The learning method according to claim 12 ,

wherein the generating the first learning model comprises performing supervised learning based on a plurality of first images categorized as the first category and a plurality of second images categorized as the second category, and

wherein the plurality of first images comprises images identified by a user as anomaly.

14 . A non-transitory storage medium storing a program that causes a computer to:

generate a first learning model for discriminating between a first category and a second category by performing supervised learning based on a first image categorized as the first category and a second image categorized as the second category;

use the first learning model to acquire a category of a first captured image as one of a plurality of categories, the plurality of categories including the first category and the second category;

receive a determination indicating whether the acquired category is either correct or incorrect;

generate a second learning model by performing supervised learning based on the first captured image and the determination, wherein the second learning model categorizes an image into the plurality of categories; and

use the second learning model to acquire a category of a second captured image as one of the plurality of categories.

15 . The non-transitory storage medium according to claim 14 , wherein the program that causes the computer to control a display apparatus to display the first captured image that is categorized as the first category by the first learning model.

16 . The non-transitory storage medium according to claim 15 , wherein the displayed first captured image is categorized as the first category with reliability equal to or higher than a predetermined level in the first learning model.

17 . The non-transitory storage medium according to claim 14 , wherein the determination is based on an input by a user via a display apparatus that displays the first captured image.

18 . The non-transitory storage medium according to claim 14 , wherein the first category is anomaly and the second category is normal.

19 . The non-transitory storage medium according to claim 18 ,

wherein the program further causes the computer to generate the first learning model by performing supervised learning based on a plurality of first images categorized as the first category and a plurality of second images categorized as the second category, and

wherein the plurality of first images comprises images identified by a user as anomaly.