Image Manager Scene Representation Graph

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing electronic systems face challenges in managing image information effectively due to increased demands for system functionality and performance, leading to resource constraints and economic inefficiencies, particularly in handling complex digital data and image categorization processes.

Innovation Solution

A system and method that employs an image manager to perform level set-based and graph-based segmentations, integrates segmentations, utilizes subscene classifiers and semantic databases for error correction, and generates image understanding and scene representation graphs to automate image understanding and classification, incorporating human knowledge and context information within a Bayesian framework.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If enhanced system functionality and performance are implemented to support advanced image categorization procedures, then image understanding capability is improved, but processing power requirements and hardware resources increase

Engineering Contradiction:

Improveimage understanding capabilityVSAvoidprocessing power requirements

Core Design Contradiction:

Adaptability or versatilityVSPower

Solution Approach 1:

The patent segments the image understanding process into distinct modules: segmentation module that divides images into regions, classification module that categorizes regions, and integration module that combines results. This modular segmentation allows each component to process only necessary data, reducing overall processing power requirements while maintaining comprehensive image understanding capability

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces a hierarchical dimension to image processing, organizing classification into multiple levels (e.g., scene level, object level, detail level). This dimensional organization enables the system to process images at different granularities, allowing users to adjust processing depth based on needs, thereby optimizing the balance between capability and processing power consumption

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If enhanced system capability to perform various advanced operations is implemented, then benefits to system user are increased, but demands on control and management of various system components increase

Engineering Contradiction:

Improvesystem capabilityVSAvoidcontrol and management complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements a universal interface layer that standardizes access to multiple image processing functions. This interface provides consistent methods for segmentation, classification, and analysis operations, allowing the system to deliver enhanced capability while maintaining simplified control and management through a single point of access regardless of which advanced operation is performed

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent introduces an intermediary management layer that sits between the user and the complex processing components. This intermediary handles the coordination and control of segmentation algorithms, classification models, and integration processes, shielding users from the underlying complexity while enabling access to advanced capabilities through simplified commands

Inventive Principle:

Principle #24Intermediary (Mediator)

3Productivity

If more hardware resources are allocated to handle complex digital data, then image processing performance is improved, but production costs and operational inefficiencies increase

Engineering Contradiction:

Improveimage processing performanceVSAvoidhardware resources

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The patent implements partial processing by allowing users to specify which portions of an image require full analysis and which can receive simplified processing. For example, critical regions identified through initial segmentation can receive intensive classification while other regions receive basic categorization, reducing overall hardware resource consumption while maintaining high performance for important areas

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent enables dynamic adjustment of processing parameters such as segmentation granularity, classification depth, and analysis thoroughness based on image characteristics and user needs. By changing these parameters, the system can optimize hardware resource utilization for different processing scenarios, achieving high performance when needed while conserving resources during routine operations

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS8873833B2System and method for effectively performing a scene representation procedure

Publication Date: 2014.10.28 SONY GROUP CORP

AI summary

A system for performing a scene representation procedure includes an image manager that processes source images from a given scene to define subscenes in the source images. The image manager creates an image understanding graph for each of the source images, and also creates a scene representation graph for each of the source images based upon the corresponding subscenes and certain image characteristics. The image manager further generates an integrated scene representation to represent all of the source images with a single representation. A processor of an electronic device controls the image manager to perform the scene representation procedure.