Enhanced Scene Detection for Depth Estimation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current image data analysis techniques face challenges in efficiently processing increased data volumes and complexities, leading to higher resource demands and economic impacts, particularly in electronic devices, which necessitate improved methods for depth estimation.

Innovation Solution

A system and method utilizing enhanced scene detection in depth estimation procedures, involving initial scene detection, radius analysis, and registration analysis to accurately classify blur images as either pillbox or Gaussian, thereby optimizing depth estimation processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If enhanced scene detection procedures are implemented to improve depth estimation accuracy, then measurement precision is improved, but device complexity increases

Engineering Contradiction:
Improvedepth estimation accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The scene detection procedure is divided into multiple independent analysis stages: initial scene detection, radius analysis, and registration analysis. Each stage processes specific features independently and contributes to the final classification decision, allowing the system to achieve high accuracy through modular processing rather than a single complex algorithm.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary scene detection to identify candidate scene types before conducting more detailed radius and registration analyses. This preliminary classification allows the system to focus computational resources on verifying specific hypotheses rather than processing all possible scene types equally, improving efficiency while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If multiple analysis procedures are performed to verify scene detection results, then reliability is improved, but productivity decreases

Engineering Contradiction:
Improveclassification accuracyVSAvoidprocessing speed
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system dynamically adjusts the depth of verification based on initial detection results. When the initial scene detection produces high-confidence results, the system may perform fewer verification steps. When confidence is lower or results are ambiguous, the system automatically performs more thorough radius and registration analyses, optimizing the balance between reliability and productivity for each specific case.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The radius analysis and registration analysis procedures provide feedback to the initial scene detection results. If the verification analyses confirm the initial classification, the system accepts the result efficiently. If discrepancies are found, the system uses the feedback to reclassify or re-analyze, ensuring high reliability while avoiding unnecessary processing when results are already correct.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP2704419B1System and method for utilizing enhanced scene detection in a depth estimation procedure
Publication Date: 2017.06.14 SONY GROUP CORP
  • EP2704419B1 patent drawingFigure 1
  • EP2704419B1 patent drawingFigure 2
  • EP2704419B1 patent drawingFigure 3

AI summary

A system for performing an enhanced scene detection procedure including a sensor device for capturing blur images of a photographic target. The blur images each correspond to a scene type that is detected from a first scene type which is typically a pillbox blur scene, and a second scene type which is typically a Gaussian scene type. A scene detector performs an initial scene detection procedure to identify a candidate scene type for the blur images. The scene detector then performs the enhanced scene detection procedure to identify a final scene type for the blur images.