Color Doppler Image Enhancement With Deep Learning at High Frame Rate

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional medical imaging systems, particularly in color Doppler imaging, face limitations in presenting sufficient feedback information during examinations, especially regarding image quality and frame rate, which are inversely related in color Doppler flow diagnostic images.

Innovation Solution

The implementation of deep learning techniques, specifically using generative adversarial networks (GANs), enhances color Doppler image quality by maintaining both image quality and frame rate through advanced processing, optimizing ensemble sizes to improve diagnostic performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional color Doppler imaging is used to maintain high frame rate, then productivity is improved, but manufacturing precision deteriorates (image quality decreases)

Engineering Contradiction:

Improveframe rateVSAvoidimage quality

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

A deep learning processing module is introduced as an intermediary between the raw color Doppler data acquisition and the final image display. This module uses trained neural networks to enhance image quality without requiring increased ensemble sizes, thereby maintaining high frame rates while improving diagnostic image quality through intelligent post-processing.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system changes the processing parameters by applying deep learning algorithms that can improve image quality metrics (such as signal-to-noise ratio and contrast) without changing the acquisition parameters (ensemble size, pulse repetition frequency). This allows decoupling the traditional inverse relationship between frame rate and image quality.

Inventive Principle:

Principle #35Parameter changes

2Manufacturing precision

If conventional color Doppler imaging increases ensemble size to improve image quality, then manufacturing precision is improved, but productivity deteriorates (frame rate decreases)

Engineering Contradiction:

Improveimage qualityVSAvoidframe rate

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

Deep learning models are pre-trained on large datasets of color Doppler images before deployment. During actual examination, these pre-trained models rapidly process the data to enhance image quality, eliminating the need for real-time iterative processing that would slow down frame rate. The preliminary training phase separates the time-consuming learning process from the real-time imaging process.

Inventive Principle:

Principle #10Preliminary action

3Device complexity

If conventional imaging presents limited feedback information, then device complexity is reduced, but measurement precision deteriorates (diagnostic information sufficiency)

Engineering Contradiction:

Improvesystem simplicityVSAvoiddiagnostic information sufficiency

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The system replaces traditional signal processing and image enhancement methods with deep learning-based processing. This substitution enables the extraction and presentation of more diagnostic information (improved measurement precision) while keeping the user interface and display complexity manageable, as the intelligence is embedded in the automated processing pipeline rather than requiring complex user controls.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

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

Deep learning techniques enable high-quality color Doppler images with improved frame rates, facilitating better user diagnosis by addressing the inverse relationship between image quality and frame rate in conventional systems.

Implementation Method 1

ultrasound imaging uses real time, non-invasive high frequency sound waves to produce ultrasound images

Methodology Applied

Scientific EffectUltrasound: Ultrasound

Implementation Method 2

processing the acquired signals includes determining Doppler effects associated with at least some of the signals

Methodology Applied

Scientific EffectDoppler effect: Doppler Effect

Data Source

PatentUS12499512B2Improving color doppler image quality using deep learning techniques

Publication Date: 2025.12.16 GE PRECISION HEALTHCARE LLC

AI summary

Systems and methods are provided for improving color Doppler image quality using deep learning techniques. In a medical imaging system, signals associated with a medical imaging technique may be acquired and processed, with the processing including determining Doppler effects associated with at least some of the signals. Medical images configured for color Doppler based examination may be generated based on the processing of the acquired signals and the determining of the Doppler effects. The medical images may be low quality color images. The medical images may be processed using at least one reference medical image corresponding to at least one of the medical images, with the at least one reference medical image being high quality color image. The processing may include applying artificial intelligence (AI) based processing, such as deep learning based modeling. Improved medical images may then be generated based on the processing of the medical images.