Ophthalmic Image Generation for Time-Specific Pseudo Contrast Imaging

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional image generation techniques fail to adequately depict contrast effects corresponding to specific contrast times, leading to limitations in acquiring and utilizing contrast-enhanced images for medical diagnosis due to adverse effects from contrast agents and radiation exposure.

Innovation Solution

An image generation apparatus and method that utilizes an image acquisition unit and an output unit to input ophthalmic examination images and contrast times into an image generation model, generating contrast-enhanced images depicting contrast effects using deep learning techniques, specifically U-Net-based models, to produce pseudo contrast images resembling fluorescein angiography examinations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If contrast agents and radiation are used to acquire contrast-enhanced images, then imaging quality and diagnostic information are improved, but harmful effects on patients increase

Engineering Contradiction:

Improveimaging qualityVSAvoidradiation exposure

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent uses a generative adversarial network (GAN) to create synthetic contrast-enhanced images that copy and replicate the appearance and diagnostic features of real contrast-enhanced images. The GAN takes non-contrast images as input and generates fake contrast-enhanced images that visually resemble real FA images, allowing diagnosis without actual contrast agents or radiation exposure.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces expensive and harmful contrast agents with computationally generated images. Instead of requiring actual contrast agents that cause adverse effects, the system uses algorithmically generated image data that provides the same diagnostic value without physical harm to the patient.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

2Productivity

If contrast agents are administered multiple times, then repeated examinations can be performed, but adverse effects accumulate

Engineering Contradiction:

Improverepeated examination capabilityVSAvoidcontrast agent accumulation

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The GAN model creates synthetic contrast-enhanced images from non-contrast images, enabling repeated virtual examinations without administering actual contrast agents. The system can generate multiple time-point images from a single non-contrast input image, simulating the appearance of sequential contrast-enhanced scans.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces the physical administration of contrast agents with a computational imaging system. Instead of injecting contrast media into the patient's body, the system uses deep learning algorithms to synthetically enhance images, substituting mechanical/chemical contrast enhancement with digital image processing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Adaptability or versatility

If conventional image generation models are used, then image conversion is possible, but accurate depiction of contrast effects at specific contrast times cannot be achieved

Engineering Contradiction:

Improveimage conversion capabilityVSAvoidcontrast effect accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent modifies the GAN model by conditioning it on specific contrast time parameters. The model takes non-contrast images and generates synthetic contrast-enhanced images at predetermined time points (e.g., 10 seconds, 30 seconds, 60 seconds after contrast agent injection). This temporal conditioning enables accurate depiction of contrast effects at specific moments, matching the precision of actual FA examinations.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The GAN model is trained beforehand using paired data of non-contrast images and their corresponding real contrast-enhanced images at various time points. This pre-training enables the model to learn and reproduce accurate contrast effect patterns, allowing it to generate precise synthetic images at any requested time point without requiring actual contrast agent administration.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20260010981A1Image generation apparatus, image generation method, and storage medium

Publication Date: 2026.01.08 CANON KK

AI summary

An image generation apparatus includes an image acquisition unit configured to acquire an ophthalmic examination image, and an output unit configured to input the ophthalmic examination image and at least one contrast time as input data of an image generation model configured to generate a contrast-enhanced image depicting a contrast effect, and thereby provide output of at least one contrast-enhanced image output as output data of the image generation model along with the at least one contrast time.