Image Processing Device Region Setting Convolution Operation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing devices face challenges in reducing operation amounts and power consumption during convolution operations using neural networks.

Innovation Solution

Incorporating a region setting section to define an operation target region based on flag data within an image region, and a convolution operation section to perform operations only in this target region using weighting coefficient data, thereby reducing unnecessary processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If convolution operation is performed on the entire image region, then processing completeness is maintained, but operation amount and power consumption increase

Engineering Contradiction:
Improveoperation amountVSAvoidprocessing completeness
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent divides the image region into multiple sub-regions and processes only the necessary sub-regions based on flag data. The region setting section segments the image into operation target regions and non-target regions, allowing the convolution operation to be performed only on relevant areas, thus reducing overall operation amount while maintaining processing completeness for important regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different processing strategies to different regions of the image. By using flag data to identify operation target regions, the system performs convolution operations only in regions where processing is necessary, while skipping non-target regions. This local differentiation reduces the total operation amount without compromising the quality of processing for critical regions.

Inventive Principle:
Principle #3Local quality

2Productivity

If convolution operation is performed on the entire image region, then processing completeness is maintained, but power consumption increases

Engineering Contradiction:
Improveprocessing efficiencyVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by stationary object

Solution Approach 1:

The patent segments the image processing task into targeted operations by using flag data to identify operation target regions. This segmentation allows the convolution operation to be performed only on necessary regions, reducing the total computational load and consequently lowering power consumption while maintaining processing efficiency for critical areas.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies partial action by performing convolution operations only on the necessary portions of the image rather than the entire image. The region setting section identifies operation target regions based on flag data, and the convolution operation section processes only these regions, avoiding unnecessary computational operations and reducing power consumption.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If region setting section is added to define operation target region, then operation amount is reduced, but device complexity increases

Engineering Contradiction:
Improveoperation amountVSAvoiddevice complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The region setting section is designed to work in conjunction with the existing convolution operation section and flag data mechanism. By integrating the region setting functionality into the existing processing pipeline, the patent achieves operation amount reduction without significantly increasing device complexity, as the region setting section leverages existing data structures and processing frameworks.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4579608A1Image processing device
Publication Date: 2025.07.02 SONY SEMICON SOLUTIONS CORP
  • EP4579608A1 patent drawingFigure 1
  • EP4579608A1 patent drawingFigure 2
  • EP4579608A1 patent drawingFigure 3~4

AI summary

An image processing device according to an embodiment of the present disclosure includes: a region setting section that is configured to set an operation target region on the basis of a piece of flag data that indicates a region to be processed and a region not to be processed in an image region indicated by a piece of image data, the operation target region being an image region that has to be processed in the image region indicated by the piece of image data; and a convolution operation section that is configured to perform a convolution operation in the operation target region on the basis of the piece of image data and a piece of first weighting coefficient data.