Ultrasonic contrast image processing method, equipment and device

By using the target length and short memory network model for curve fitting of respiratory motion compensation and feedforward neural network model in ultrasound contrast image processing, the problem of inaccurate signal intensity of ultrasound contrast image is solved, and more accurate lesion diagnosis is achieved.

CN119700174AInactive Publication Date: 2025-03-28QINGDAO HISENSE MEDICAL EQUIP
8 Cites 0 Cited by

Patent Information

Application Number
CN202311253815.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-09-26
Publication Date
2025-03-28
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

During the acquisition of ultrasound contrast image, factors such as respiratory movement and heart rate pulsation affect the accuracy of signal intensity, resulting in inaccurate fitting curves and inability to obtain reliable evaluation parameters.

Method used

The target length and short memory network model is used to compensate each frame of image data in the ultrasound contrast image, and the compensated data is curved with the feedforward neural network model to determine multiple ultrasound contrast parameters.

Benefits of technology

Through precise respiratory motion compensation and curve fitting, the accuracy of ultrasound contrast image quantitative analysis is improved, and the degree of lesions in the target site can be described more accurately, thereby improving the accuracy of lesion diagnosis.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of image processing, discloses an ultrasonic contrast image processing method, equipment and device, and aims to provide a more accurate quantitative analysis scheme for an ultrasonic contrast image. The method comprises the following steps: acquiring an ultrasonic contrast image; a target long-short memory network model is used to carry out breathing motion compensation on each frame of image data in the ultrasound contrast image to obtain compensated multi-frame image data, the target long-short memory network model comprises a compensation layer, and the compensation layer comprises at least two Bi-LSTM layers connected in sequence; a feedforward neural network model is used for conducting curve fitting on the compensated multi-frame image data, and a curve is used for describing the relation between the ultrasound contrast time and the ultrasound contrast signal intensity; based on the obtained curve, multiple ultrasound contrast parameters are determined, and the ultrasound contrast parameters are used for describing the lesion degree of the target part corresponding to the ultrasound contrast image.
Need to check novelty before this filing date? Find Prior Art