Dual-Resolution Image Recognition Circuit for Speed-Accuracy Trade-off

Solution Overview

Problem

Existing image recognition techniques face a trade-off between processing speed and accuracy, with high-resolution images enabling detailed recognition but at slower speeds, and low-resolution images allowing faster processing but only simple recognition.

Innovation Solution

An image recognition device that separates the image recognition process into two stages, using low-resolution images for initial high-speed simple recognition and high-resolution images for subsequent detailed recognition, with charge storage circuits and read circuits configured to output images of different resolutions based on exposure periods and pixel averaging.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If high resolution image is used for feature amount calculation, then recognition accuracy is improved, but processing speed is reduced

Engineering Contradiction:

Improverecognition accuracyVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent divides the image recognition process into two distinct stages: a first stage using low resolution images for initial defect candidate detection, and a second stage using high resolution images for detailed verification. This segmentation allows the system to leverage the speed advantage of low resolution processing while maintaining the accuracy benefit of high resolution processing for critical evaluations.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary image processing by capturing and processing a low resolution image before the high resolution image. This preliminary action identifies potential defect candidates early in the process, allowing the subsequent high resolution processing to focus only on regions of interest, thereby improving overall processing efficiency.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If low resolution image is used for feature amount calculation, then processing speed is increased, but only simple recognition is possible

Engineering Contradiction:

Improveprocessing speedVSAvoidrecognition accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent divides the image recognition process into two distinct stages: a first stage using low resolution images for initial defect candidate detection, and a second stage using high resolution images for detailed verification. This segmentation allows the system to leverage the speed advantage of low resolution processing while maintaining the accuracy benefit of high resolution processing for critical evaluations.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary image processing by capturing and processing a low resolution image before the high resolution image. This preliminary action identifies potential defect candidates early in the process, allowing the subsequent high resolution processing to focus only on regions of interest, thereby improving overall processing efficiency.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If two stage image recognition process is implemented, then both high accuracy and high speed recognition are realized, but device complexity is increased

Engineering Contradiction:

Improverecognition accuracyVSAvoidprocess complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent merges the low resolution and high resolution image processing into a unified two-stage workflow where the first stage outputs feed into the second stage. By integrating these processes and using the first stage results to guide the second stage, the system achieves both speed and accuracy while managing complexity through coordinated operation of the combined processes.

Inventive Principle:

Principle #5Merging (Combining)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enables both high-accuracy and high-speed image recognition by leveraging the benefits of both low- and high-resolution images without the need for additional image resolution conversion, reducing processing time and power consumption while maintaining miniaturization.

Implementation Method 1

signal charges generated by photoelectric conversion sections provided in each of a plurality of pixels

Methodology Applied

Scientific EffectPhotoelectric conversion: Photoelectric Effect

Data Source

PatentUS10089548B2Image recognition device and image recognition method

Publication Date: 2018.10.02 OLYMPUS CORPORATION(JP)

AI summary

An image recognition device includes: a plurality of first charge storage circuits that store signal charges generated by photoelectric conversion sections; a plurality of second charge storage circuits that store signal charges generated by the photoelectric conversion sections; a first charge read circuit section that reads a pixel signal and outputs an image as a first image; a second charge read circuit section that reads a pixel signal and outputs an image as a second image; a read circuit selection section that selects one of the first charge read circuit section and the second charge read circuit section; and a feature amount determination section, wherein the feature amount determination section determines a detection target subject according to a feature amount of a subject in the second image, and whether to perform the determination for a subject in the first image is determined based on the determination result.