US-12626529-B2 - Character determination system and character determination method

Abstract

An object of the present invention is to provide a character determination system and the like which determine a character of an animal with a simple method. The character determination system includes: receiving means which receives an input of an image of an animal; and character determining means which determines, using a trained model, a character of the animal from the image of the animal inputted to the receiving means, wherein the trained model is a trained model having learned a relationship between an image of an animal and a character of the animal.

Inventors

• Tsuguna Inagaki
• Hiroyuki Takahashi
• Yoshie OGOSHI

Assignees

• ANICOM HOLDINGS, INC.

Dates

Publication Date

20260512

Application Date

20221026

Priority Date

20211029

Claims (18)

1. 1 . A character determination system, comprising: receiving means which receives an input of an image of an animal other than a human; and character determining means which determines, using a machine learning or a deep learning trained model, a character of the animal from the image of the animal inputted to the receiving means, and a processor configured to: determine behavioral characteristics of the animal; and determine the character based on a strength of a tendency for each of the behavioral characteristics that are determined; wherein the character includes friendliness, and wherein the trained model has been subjected to learning using, as training data, an image of the animal and a label related to the character of one or more specific breeds to learn a relationship between the image of the animal other than a human and the character of the animal.
2. 2 . The character determination system according to claim 1 , wherein the character includes behavioral characteristics defined in C-barq (Canine Behavioral Assessment and Research Questionnaire).
3. 3 . The character determination system according to claim 1 , wherein the trained model of the character determining means is the trained model which performs learning using the image of an animal other than the human and the label related to the character of the animal as training data and which receives the image of an animal as input and outputs the character determination of the animal.
4. 4 . The character determination system according to claim 1 , wherein the image of the animal received by the receiving means is an image obtained by photographing a face of the animal from front.
5. 5 . A character determination system, comprising: receiving means which receives an input of an image of an animal other than a human; and character determining means which determines, using a machine learning or deep learning trained model, a character of the animal from the image of the animal inputted to the receiving means, and a processor configured to: determine behavioral characteristics of the animal; and determine the character based on a strength of a tendency for each of the behavioral characteristics that are determined; wherein the trained model is a trained model having learned a relationship between an image of an animal other than a human and a character of the animal, and the character determination system further comprises breed determining means which determines, using the machine learning or deep learning trained model for breed determination, a breed of an animal from an image of the animal inputted to the receiving means.
6. 6 . The character determination system according to claim 5 , wherein the character determining means is provided with a plurality of trained models, and the character determining means determines, using one of the plurality of trained models corresponding to the determination result of the breed determining means, the character of the animal in the input image.
7. 7 . The character determination system according to claim 6 , wherein the trained model is provided in plurality in accordance with a category into which a breed of an animal is categorized based an average weight of an adult of the breed.
8. 8 . The character determination system according to claim 5 , wherein the trained model of the character determining means is a trained model which performs learning using an image of an animal other than a human and a label related to a character of the animal as training data and which receives an image of an animal as input and outputs a character determination of the animal.
9. 9 . The character determination system according to claim 5 , wherein an image of an animal received by the receiving means is an image obtained by photographing a face of the animal from front.
10. 10 . The character determination system according to claim 5 , wherein the character includes behavioral characteristics defined in C-barq (Canine Behavioral Assessment and Research Questionnaire).
11. 11 . The character determination system according to claim 5 , wherein the trained model has been subjected to learning using, as training data, an image of an animal and a label related to a character of the one or more specific breeds.
12. 12 . The character determination system according to claim 3 , wherein the trained model has been subjected to learning using, as training data, an image of an animal and a label related to a character of the one or more specific breeds.
13. 13 . The character determination system according to claim 4 , wherein the trained model has been subjected to learning using, as training data, an image of an animal and a label related to a character of the one or more specific breeds.
14. 14 . The character determination system according to claim 6 , wherein the trained model has been subjected to learning using, as training data, an image of an animal and a label related to a character of the one or more specific breeds.
15. 15 . A generation method of a machine learning or deep learning trained model which determines a character of an animal other than a human from an image of the animal, the generation method including: inputting an image of an animal other than a human and a label related to a character of the animal to a computer as training data and causing an artificial intelligence to learn the training data, wherein the character includes friendliness.
16. 16 . A character determination method, comprising: a step of preparing an image of an animal other than a human; and a step of inputting the image to a machine learning or deep learning trained model and outputting a character determination of the animal from the inputted image of the animal by a computer using the machine learning or deep learning trained model, wherein the machine learning or deep learning trained model is a trained model having learned a relationship between an image of an animal other than a human and a character of the animal, and the character includes friendliness.
17. 17 . The case determination method according to claim 16 , wherein the trained model is a trained model which performs learning using an image of an animal other than a human and a label related to a character of the animal as training data and which receives an image of an animal as input and outputs a character determination of the animal.
18. 18 . A character determination method, comprising: a step of preparing an image of an animal other than a human; and a step of breed determining a breed of an animal from an image of the animal inputted to the receiving means by a computer using a machine learning or deep learning trained model for breed determination; and a step of outputting a character determination of the animal from the inputted image of the animal by the computer using the machine learning or deep learning trained model, where the machine learning or deep learning trained model is a trained model having learned a relationship between an image of an animal other than a human and a character of the animal.

Description

TECHNICAL FIELD The present invention relates to a character determination system and a character determination method and, more specifically, to a character determination system and a character determination method which provide a determination result of a character of an animal from an image of the animal. BACKGROUND ART Pets such as dogs, cats, and rabbits are irreplaceable to humans and various pet-oriented services are being promoted and developed. Each pet is an individual with a its own personality which differ from one individual to the next even in a same species or a same breed. In addition, the character of an animal is correlated to compatibility of the animal with a pet owner and how easily the animal can be reared. For example, rearing becomes difficult if an animal has an aggressive character but an animal with an outgoing character can be readily taken to places where other animals are present. If a character of an animal can be known by a simple method, for example, a customer about to purchase the animal for rearing will find that such information is useful. While an animal is generally purchased at pet shops or through breeders, since a decision to purchase an animal is usually made in a short amount of time at a pet shop, there is not enough time to assess a character of the animal. A similar situation occurs when purchasing an animal from a breeder. Recently, information on animals being sold by pet shops and breeders including a breed, a headshot, an age in months, and a price of the animals are made available through the Internet and prospective purchasers increasingly visit an actual pet shop or a breeder where an individual of interest is being sold only after due deliberation based on the information available on the Internet. In such a case, making information on a character of an individual available on the Internet in addition to information such as the breed, the headshot, the age in months, and the price of the animal is expected to help a pet owner with making a selection and prevent a mismatch between the owner and the animal. Even a pet owner already rearing an animal will conceivably benefit from gaining knowledge of a character of the individual being reared. Since animals are unable to speak, an animal's actual character may differ from the character of the animal as perceived by the owner. In such a case, conceivably, there is a greater need for comprehending a tendency of a character of an animal by a simple method rather than comprehending the character based on a detailed analysis. Accordingly, there is a demand for means which determines a character of an animal with a simple method. While Patent Literature 1 discloses a character analysis apparatus and a character analysis program which aim to more accurately analyze a character of a user viewing a web page using information on behavioral characteristics of the user, there is no disclosure on character determination of animals. In addition, while Non Patent Literature 1 discloses an attribute determination method of a poster from an SNS profile image using deep learning, there is no disclosure on character determination of animals. RELATED ART DOCUMENTS Patent Document Patent Literature 1: JP2018-195099 Non-Patent Document Non Patent Literature 1: Yuta Yamashita, Junichiro Mori, “Predicting Demographics and Personalities from Profile Images of Social Network Users using Convolutional Deep Neural Network”, The 30th Annual Conference of the Japanese Society for Artificial Intelligence, 2016 SUMMARY OF THE INVENTION Problems to be Solved by the Invention In consideration thereof, an object of the present invention is to provide a character determination system and a character determination method which determine a character of an animal with a simple method. Means for Solving the Problems Through extensive studies carried out in order to solve the problem described above, the present inventors have found that the problem can be solved by a trained model having learned a relationship between an image of an animal and a character of the animal, culminating in completing the present invention. Specifically, the present invention provides [1] to [11] below. [1] A character determination system, including: receiving means which receives an input of an image of an animal; and character determining means which determines, using a trained model, a character of the animal from the image of the animal inputted to the receiving means, wherein the trained model is a trained model having learned a relationship between an image of an animal and a character of the animal. [2] A character determination system, including: receiving means which receives an input of an image of an animal; and character determining means which determines, using a trained model, a character of the animal from the image of the animal inputted to the receiving means, wherein the trained model is a trained model which performs learning using an image of an a