Information processing apparatus, information processing method, and information processing system

By releasing only for subsequent networks under the use environment of the AI ​​camera, using intermediate feature maps to select or generate subsequent networks, the problems of long time required for AI model adaptation and personal information leakage are solved, and the effect of shortening time and reducing data volume is achieved.

CN119768798APending Publication Date: 2025-04-04SONY SEMICON SOLUTIONS CORP
1 Cites 0 Cited by

Patent Information

Application Number
CN202380055835.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-10-24
Filing Date
2023-07-20
Publication Date
2025-04-04

AI Technical Summary

Technical Problem

Under the use environment of AI cameras, the long time required to relearn the entire AI model leads to an increase in the time required for the model to adapt and requires a large amount of input data, increasing the risk of communication data and personal information leakage.

Method used

By obtaining the intermediate feature map in the predetermined intermediate layer of the artificial intelligence model, selecting or generating a subsequent network from multiple candidate subsequent networks, re-learning is performed only for subsequent networks, reducing the number of learnings and the required input data.

Benefits of technology

It shortens the time required for AI models to adapt to environmental conditions, reduces the amount of communication data, and reduces the possibility of personal information leakage.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

An information processing device according to the present technology comprises: a subsequent network acquisition unit to which an intermediate feature map is input from an external device, the map being acquired in a prescribed intermediate layer of an artificial intelligence model, and a subsequent network acquisition unit that selects one subsequent network from a plurality of candidates for the subsequent network on the basis of the intermediate feature map, and transmits the selected subsequent network to an artificial intelligence model that includes a neural network and uses detection data from the sensor device as input data when the input data is transferred to the artificial intelligence model, the follow-up network is a network of a follow-up stage of a specified middle layer in an artificial intelligence model, or a follow-up network is generated on the basis of a training network serving as a basis and on the basis of the middle feature map; and a transmission processing unit that performs a process of transmitting the setting data of the subsequent network selected or generated by the subsequent network acquisition unit to the outside.
Need to check novelty before this filing date? Find Prior Art