Image State Detection for Optimal Object Trimming

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing methods fail to ensure optimal trimming of objects in images when the angular field or orientation of the imaging device changes, leading to extraction of unintended objects.

Innovation Solution

An image processing device with an image state detection unit and a trimming-decision processing unit that detects changes in the image state and decides whether to perform trimming based on the detected state, preventing extraction of objects different from the user's intention.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If trimming is performed based on focus position within angular field, then object extraction is automated, but wrong object may be extracted when angular field or orientation changes

Engineering Contradiction:
Improveautomatic object extractionVSAvoidaccuracy of object extraction
Core Design Contradiction:
Extent of automationVSReliability

Solution Approach 1:

The patent applies preliminary action by performing trimming decision processing immediately after focus determination processing, before any angular field or orientation changes can occur. The system detects the image state right after focus lock and makes the trimming decision based on this initial state, preventing subsequent changes from affecting the extraction accuracy. This timing strategy ensures that the trimming is based on the intended focus object rather than objects that may appear after device movement.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback by introducing an image state detection unit that continuously monitors changes in the captured image state. When the detection unit identifies that the image state has changed (due to angular field or orientation changes), it feeds this information back to the trimming decision processing unit, which then adjusts or cancels the trimming operation. This feedback mechanism ensures that trimming decisions are dynamically adjusted based on actual image state changes, maintaining extraction accuracy.

Inventive Principle:
Principle #23Feedback

2Productivity

If trimming decision is made immediately after focus lock, then processing speed is improved, but changes in image state may cause incorrect extraction

Engineering Contradiction:
Improveprocessing speedVSAvoidaccuracy of object extraction
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system performs the trimming decision processing as a preliminary action immediately after focus lock is achieved, capturing the image state at this critical moment. This timing ensures that the trimming decision is made based on the intended focus object before any device movement or orientation changes can occur, thereby maintaining both processing speed and extraction accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The image state detection unit continuously monitors for changes in the captured image state and provides feedback to the trimming decision processing unit. If changes are detected after the initial trimming decision, the system can adjust or cancel the trimming operation, ensuring accuracy is maintained without significantly delaying the overall processing speed.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS10178298B2Image processing device, image processing method, and recording medium for optimal trimming of a captured image
Publication Date: 2019.01.08 SONY GROUP CORP
  • US10178298B2 patent drawing
  • US10178298B2 patent drawing
  • US10178298B2 patent drawing

AI summary

Provided is an image processing device including an image state detection unit which detects a state of a captured image after a processing for determining a captured image has been performed, and a trimming-decision processing unit which performs a trimming decision of the captured image based on a state of the captured image.