CN-122024189-A - Server and information processing apparatus

Abstract

The present invention relates to a server and an information processing apparatus. The server includes a storage unit for storing an AI model for detecting an identification object from a captured image, and a control unit for outputting information indicating that the identification object has changed when a condition is satisfied that a place where the identification object that has been able to be identified has not been identified exceeds a 1 st threshold.

Inventors

• MURASHITA KIMITAKA

Assignees

• 丰田自动车株式会社
• 株式会社电装天

Dates

Publication Date

20260512

Application Date

20251023

Priority Date

20241112

Claims (6)

1. 1. A server is provided with: a storage unit for storing an AI model for detecting the identification object from the captured image, and The control unit outputs information indicating that the identification object has changed when a condition is satisfied that the point where the identification object that has been unable to be identified exceeds the 1 st threshold.
2. 2. The server of claim 1, wherein, When the condition is satisfied, the control unit collects an image including the changed recognition target as training data, and updates the AI model.
3. 3. The server according to claim 1 or 2, wherein, The control unit creates a map in which the position of the identification object is mapped, When a condition that a place where an identification target cannot be identified is not more than a 2 nd threshold is satisfied, the control unit deletes map data of the place from the map, wherein the 2 nd threshold is not more than the 1 st threshold.
4. 4. An information processing apparatus, wherein the information processing apparatus comprises: a storage unit that stores the AI model received from the server; a control unit for inputting a captured image captured by an in-vehicle camera to the AI model, performing recognition of a recognition object, and determining whether or not a capturing condition of the captured image is the same as that in the past when the recognition object can be recognized, if the recognition object cannot be recognized from the captured image And a communication unit configured to transmit, to the server, the photographing condition, the photographed image in which the identification object cannot be identified, and information notifying that the identification object cannot be identified under the photographing condition in which the identification object can be identified in the past, when the identification object cannot be identified from the photographed image and the photographing condition is the same as that when the identification object can be identified in the past.
5. 5. The information processing apparatus according to claim 4, wherein, When the identification object can be identified from the captured image, the communication unit transmits the captured condition of the captured image and the captured image in which the identification object can be identified to the server.
6. 6. The information processing apparatus according to claim 5, wherein, The shooting condition is information including a position of a vehicle on which the in-vehicle camera is mounted and an orientation of the vehicle.

Description

Server and information processing apparatus Technical Field The present disclosure relates to a server and an information processing apparatus. Background Conventionally, a technique for screening learning data contributing to improvement of the reasoning result of a model with a low load is known. For example, japanese patent application laid-open No. 2024-073155 discloses a learning data screening apparatus including an information applying section for applying, as meta information, environmental information of a storage area, which affects ease of recognition of an article, fine information related to resolution of an image, and/or inference information based on output of a model for recognizing an article, which are included in an image obtained by imaging the storage area for storing the article, to the image, and an image screening section for screening an image of a storage section storing a plurality of images to which meta information is applied, based on the meta information. Disclosure of Invention The technique described in japanese patent application laid-open No. 2024-073155 screens learning data according to the difficulty in identifying an article, but the condition is that an article can be identified, and the condition that an identifiable article becomes unrecognizable is not considered. For example, in the case where a stored article is put into a box, the design of the box is changed, it becomes impossible to identify the article. When such a design change or the like occurs, the operator needs to visually confirm that the design change has occurred, collect image data of a new design, create learning data, and update (relearn) the model. The present disclosure, which has been made in view of the above circumstances, has an object to be able to automatically detect a change in an identification object accompanying an environmental change. The server according to an embodiment of the present disclosure includes a storage unit that stores an AI model for detecting an identification object from a captured image, and a control unit that outputs information indicating that the identification object has changed when a condition that a place where the identification object that has been able to be identified has not been identified exceeds a1 st threshold is satisfied. An information processing device according to an embodiment of the present disclosure includes a storage unit that stores an AI model received from a server, a control unit that inputs a captured image captured by an in-vehicle camera to the AI model, performs recognition of a recognition object, determines whether or not a capturing condition of the captured image is the same as that when the recognition object was previously recognized, if the recognition object was not recognized from the captured image, and transmits, to the server, the capturing condition, the captured image in which the recognition object was not recognized, and information notifying that the recognition object was not recognized under the capturing condition in which the recognition object was previously recognized, if the recognition object was not recognized from the captured image, and the capturing condition was the same as that when the recognition object was previously recognized. According to an embodiment of the present disclosure, a change in an identification object accompanying an environmental change can be automatically detected. Drawings Features, advantages, and technical and industrial significance of exemplary embodiments of the present invention will be described below with reference to the accompanying drawings, in which like reference numerals denote like elements, and in which: Fig. 1 is a diagram showing a configuration example of an image recognition system according to an embodiment; FIG. 2 is a flowchart showing an example of processing steps of the vehicle according to an embodiment, and Fig. 3 is a flowchart showing an example of processing steps of the server according to the embodiment. Detailed Description Structure of image recognition system Referring to fig. 1, the structure of an image recognition system according to an embodiment will be described. The image recognition system 1 includes a plurality of vehicles 10 and a server 20. In fig. 1, only 1 vehicle is shown for convenience. The plurality of vehicles 10 and the server 20 are connected to be able to communicate with each other via a network 30 including the internet. The server 20 may be of a cloud type or a local deployment type. The vehicle 10 includes an information processing device 11, an in-vehicle camera 12, and a sensor 13. The in-vehicle camera 12 is a camera mounted on the vehicle 10, and may be mounted in plural. The in-vehicle camera 12 is connected to the information processing device 11. The in-vehicle camera 12 may be integrated with the information processing device 11. The in-vehicle camera 12 has an imaging element, and captures an image of the pe