Neural Network Diagnosis via Sample Data Extraction

Solution Overview

Problem

Conventional image diagnosis methods require doctors to manually check entire radiographs for abnormalities, leading to a heavy workload and increased complexity, with existing neural network methods requiring extraction of regions of interest before diagnosis, which complicates the system and increases arithmetic processing load.

Innovation Solution

A diagnosis processing device that uses a neural network to diagnose abnormalities by digitalizing images, sampling data from these images using a predetermined method, and inputting the sampled data to a learned neural network to determine the substantive feature of the abnormality, eliminating the need for extracting regions of interest and reducing the amount of data required for processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If regions of interest are extracted from radiograph data before inputting to neural network, then detection accuracy is improved, but system complexity and arithmetic processing load increase

Engineering Contradiction:

Improvedetection accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the necessary sample data rows from the entire radiograph data, rather than extracting regions of interest. This selective extraction of data rows reduces the data volume input to the neural network while maintaining detection accuracy, thereby reducing system complexity and arithmetic processing load

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the radiograph data into multiple sample data rows and selects specific rows for neural network input. This segmentation approach allows the system to process a manageable subset of data that retains the essential diagnostic information, balancing accuracy with reduced computational complexity

Inventive Principle:

Principle #1Segmentation

2Measurement precision

If regions of interest are extracted from radiograph data before inputting to neural network, then detection accuracy is improved, but arithmetic processing load increases

Engineering Contradiction:

Improvedetection accuracyVSAvoidarithmetic processing load

Core Design Contradiction:

Measurement precisionVSPower

Solution Approach 1:

The patent extracts only the necessary sample data rows from the entire radiograph data, rather than extracting regions of interest. This selective extraction of data rows reduces the data volume input to the neural network while maintaining detection accuracy, thereby reducing system complexity and arithmetic processing load

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent uses a partial subset of radiograph data (specific sample data rows) rather than processing the complete dataset or extracted regions of interest. This partial action approach provides sufficient diagnostic information while significantly reducing the arithmetic processing load on the neural network

Inventive Principle:

Principle #16Partial or excessive action

3Device complexity

If entire radiograph data is used for neural network input, then system simplicity is maintained, but detection accuracy decreases

Engineering Contradiction:

Improvesystem simplicityVSAvoiddetection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent extracts only the necessary sample data rows from the entire radiograph data, rather than extracting regions of interest. This selective extraction of data rows reduces the data volume input to the neural network while maintaining detection accuracy, thereby reducing system complexity and arithmetic processing load

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent changes the parameter of data selection from using entire radiograph data or extracted regions of interest to using specifically selected sample data rows. This parameter change optimizes the balance between system simplicity and detection accuracy by selecting data rows that contain the most diagnostic information

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS8798345B2Diagnosis processing device, diagnosis processing system, diagnosis processing method, diagnosis processing program and computer-readable recording medium, and classification processing device

Publication Date: 2014.08.05 SHARP KK

AI summary

A diagnosis processing device is provided in which diagnosis is realizable by a simple arrangement. A diagnosis processing device (1) of the present invention includes: a learning pattern creating section (10a) for creating a learning pattern by sampling data from a learning image in which abnormality information indicating a substantive feature of abnormality of a target is pre-known; a learning processing section (12) for causing a neural network (17) to learn, by using learning patterns; a diagnostic pattern creating section (10b) for creating a diagnostic pattern by sampling data from a diagnostic image in which abnormality information is unknown; a determination processing section (18) for determining a substantive feature of the abnormality of the target indicated in the abnormality information in the diagnostic image, based on an output value outputted, in response to an input of the diagnostic pattern, from a learned neural network (17) which is a neural network subjected to learning.