Endoscope Image Color Tone Normalization for Device-Agnostic Lesion Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing endoscope devices capture images with varying color tones and qualities, limiting the use of accumulated images for accurate lesion diagnosis, and frequent device replacements necessitate retraining, creating a vicious cycle of reduced accuracy.
Innovation Solution
An endoscope diagnosis program and device that perform color tone correction and image quality enhancement on images from different devices, using a trained model generated through machine training of corrected images, allowing for unified diagnosis across devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If machine training is performed using all endoscope images from different devices, then the quantity of training data increases, but the diagnosis accuracy deteriorates due to varying color tones and image qualities
Solution Approach 1:
The patent applies parameter changes by performing color tone correction on training images from different endoscope devices to unify their color characteristics. This transforms images with varying color tones and qualities into a standardized format, allowing the trained model to learn from a large quantity of images without accuracy degradation due to device-specific variations
2Measurement precision
If machine training is performed for each device type separately, then the diagnosis accuracy improves for that device, but the quantity of usable images decreases and retraining is required upon device replacement
Solution Approach 1:
The patent implements universality by creating a single trained model that can accurately diagnose lesions across multiple endoscope device types. Through color tone correction and unified training on images from various devices, the model becomes device-agnostic, eliminating the need for separate training per device type and enabling immediate use upon device replacement
3Adaptability or versatility
If device replacement occurs, then access to newer technology is achieved, but the training process must be restarted causing time loss and reduced accuracy
Solution Approach 1:
The patent applies preliminary action by performing color tone correction and creating a unified trained model in advance that accounts for variations across different device types. This preparatory work ensures that when device replacement occurs, the model is already optimized to handle images from the new device without requiring retraining, thus eliminating time loss
Data Source
Figure 1
Figure 2
Figure 3
AI summary
[Problem] To provide an endoscope diagnosis program, an endoscope diagnosis device, a control method for an endoscope diagnosis device, and a program for generating an endoscope diagnosis trained model which each make it possible to use endoscope images that have been accumulated to date to diagnose the presence or absence of a lesion, without limitations on the type of endoscope device. [Solution] A computer is made to function as: a diagnostic-use image acquisition unit 66 that acquires a diagnostic-use image of a site that is subject to diagnosis which has been captured with an endoscope device 20; a color tone correction unit 67 that performs color tone correction on the diagnostic-use image in accordance with a reference color tone, which is a color tone of an image of a site subject to diagnosis having been captured in advance with the endoscope device; and a lesion presence/absence diagnosis unit 68 that inputs the corrected diagnostic-use image into a trained model and that performs a process for diagnosing the presence or absence of a lesion from a result output by the trained model, said trained model having being generated by machine learning of a plurality of training images which have been subjected to color tone correction with the reference color tone.