Automatic Image Capture Learning From User Deletion Feedback

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Automatic image capturing systems often capture scenes that do not match a user's taste, as they rely on periodic shooting intervals without user instruction, leading to unintended moments being recorded.

Innovation Solution

An image processing apparatus and method that determines image data for learning by analyzing user deletion instructions, using a neural network to identify images that do not match a user's taste, and adjusting automatic image capturing settings accordingly.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If automatic image capturing is performed at a certain time interval, then the quantity of captured images increases, but the quality of captured images (matching user's taste) deteriorates

Engineering Contradiction:

Improvequantity of captured imagesVSAvoidquality of captured images

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The system uses user deletion operations as feedback signals to learn and improve automatic image capturing quality. When users delete captured images, the system analyzes these deletion patterns to understand what types of images do not match user taste, then adjusts future capturing decisions based on this learned feedback, creating a continuous improvement loop.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs self-learning by automatically analyzing user deletion operations without requiring explicit user instructions or operations. The deletion operations themselves serve as the learning input, allowing the system to autonomously improve its image capturing quality assessment based on actual user behavior patterns.

Inventive Principle:

Principle #25Self-service

2Measurement precision

If machine learning is trained using only images that match user's taste, then the learning effectiveness improves, but the amount of available training data decreases

Engineering Contradiction:

Improvelearning effectivenessVSAvoidamount of training data

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The system converts user deletion operations, which initially seem to reduce available data, into valuable training opportunities. By analyzing why users delete images, the system transforms negative feedback (deletions) into positive learning data that helps improve future capturing decisions, thereby increasing both the quantity and quality of effective training data.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Data Source

PatentUS11438501B2Image processing apparatus, and control method, and storage medium thereof

Publication Date: 2022.09.06 CANON KK

AI summary

An image processing apparatus comprises a determination unit configured to determine that image data stored in a storage device is image data that is to be used for learning a situation in which an image capturing device is to perform automatic image capturing, if an instruction to delete the image data stored in the storage device is given by a user, and the image data satisfies a predetermined condition.