Image Capturing Element Local Feature Extraction for Privacy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image capturing systems face risks of personal information leakage when transmitting captured images to the cloud for feature extraction, compromising communication safety and image recognition performance.

Innovation Solution

An image capturing element and device that captures images and extracts predetermined feature amounts locally, encoding and transmitting these instead of raw images, thereby preventing data leakage and enhancing privacy protection while maintaining accurate safety determination through image recognition.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If captured images are transmitted to the cloud for feature extraction, then image recognition performance is improved, but communication safety deteriorates due to personal information leakage risks

Engineering Contradiction:
Improveimage recognition performanceVSAvoidcommunication safety
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent extracts only the essential feature amounts from captured images and transmits these extracted features to the cloud, rather than transmitting the entire captured images. This extraction process removes unnecessary personal information while preserving the essential characteristics needed for image recognition, thereby maintaining recognition performance while improving communication safety.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the image data processing into two parts: local feature extraction performed by the image capturing device, and cloud-based analysis performed by the server. This segmentation allows the device to process and anonymize data locally before transmission, reducing the risk of personal information leakage while enabling comprehensive cloud-based image recognition analysis.

Inventive Principle:
Principle #1Segmentation

2Loss of information

If captured images are transmitted instead of feature amounts, then complete image data is available for analysis, but data security and privacy protection deteriorate

Engineering Contradiction:
Improveimage data completenessVSAvoidpersonal information leakage
Core Design Contradiction:
Loss of informationVSObject-affected harmful factors

Solution Approach 1:

The patent extracts only the essential feature amounts from captured images and transmits these extracted features to the cloud, rather than transmitting the entire captured images. This extraction process removes unnecessary personal information while preserving the essential characteristics needed for image recognition, thereby maintaining recognition performance while improving communication safety.

Inventive Principle:
Principle #2Taking out (Extraction)

3Reliability

If feature amounts are extracted and transmitted locally, then communication safety is improved, but image recognition accuracy may deteriorate

Engineering Contradiction:
Improvecommunication safetyVSAvoidimage recognition accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The patent extracts only the essential feature amounts from captured images and transmits these extracted features to the cloud, rather than transmitting the entire captured images. This extraction process removes unnecessary personal information while preserving the essential characteristics needed for image recognition, thereby maintaining recognition performance while improving communication safety.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the image data processing into two parts: local feature extraction performed by the image capturing device, and cloud-based analysis performed by the server. This segmentation allows the device to process and anonymize data locally before transmission, reducing the risk of personal information leakage while enabling comprehensive cloud-based image recognition analysis.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS11956526B2Image capturing element, image capturing device and method
Publication Date: 2024.04.09 SONY GROUP CORP
  • US11956526B2 patent drawing
  • US11956526B2 patent drawing
  • US11956526B2 patent drawing

AI summary

The present disclosure relates to an image capturing element, an image capturing device, and a method that are capable of preventing a reduction in communication safety.By capturing an image of a subject, and generating a captured image, a predetermined feature amount to be output to the outside is extracted from the generated captured image. For example, an image capturing section that captures an image of a subject to generate a captured image, and a feature amount extracting section that extracts, from the captured image generated by the image capturing section, a predetermined feature amount to be output to the outside, are packaged and are formed in respective substrates different from each other. The present disclosure can be applied to, for example, an image capturing element, an image capturing device, an image processing device, a transmission device, or a communication device.