Absolute flow velocity quantification method for laser speckle blood flow imaging and related device

By employing physical enhancement preprocessing and domain adversarial regression networks, this study addresses the limitation of traditional laser speckle blood flow imaging technology in providing full-field absolute flow velocity measurement during cardiac surgery. It achieves high-precision blood flow monitoring, adapts to complex physiological environments, and provides a new method for quantitative assessment of absolute flow velocity in cardiac surgery.

CN122156171APending Publication Date: 2026-06-05TECHNICAL INST OF PHYSICS & CHEMISTRY - CHINESE ACAD OF SCI
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
TECHNICAL INST OF PHYSICS & CHEMISTRY - CHINESE ACAD OF SCI
Filing Date
2026-03-13
Publication Date
2026-06-05

AI Technical Summary

Technical Problem

Traditional laser speckle flow imaging technology cannot provide absolute flow velocity measurement across the entire field of view in cardiac surgery, and deep learning models have insufficient generalization ability in complex physiological backgrounds, resulting in unstable measurement results.

Method used

By employing physical enhancement preprocessing and domain adversarial regression networks, a dual-channel spatiotemporal tensor is constructed. Domain-invariant features independent of tissue background are extracted through the domain adversarial regression network. Combined with a few-sample transfer learning strategy, a leapfrog quantitative measurement from relative blood flow index to absolute flow velocity is achieved.

Benefits of technology

It achieves full-field, non-contact, and high-precision blood flow monitoring, overcomes the insufficient quantitative capabilities of traditional methods, adapts to complex physiological environments, and provides a new means for quantitative assessment of absolute flow velocity in cardiac surgery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122156171A_ABST
    Figure CN122156171A_ABST
Patent Text Reader

Abstract

The application provides an absolute flow velocity quantification method and device for laser speckle blood flow imaging, an electronic equipment and a storage medium. The method comprises the following steps: acquiring an original speckle image sequence of a sample to be measured; performing physical enhancement preprocessing on the original speckle image sequence to construct a double-channel space-time tensor containing spatial structure and time fluctuation information; inputting the double-channel space-time tensor into a trained domain adversarial regression network to extract domain-invariant features irrelevant to tissue background; and based on the domain-invariant features, an absolute blood flow velocity value of the sample to be measured is calculated by regression. The combination of physical enhancement preprocessing and the domain adversarial regression network realizes a leapfrog quantitative measurement of laser speckle blood flow imaging from a relative blood flow index to an absolute flow velocity. Meanwhile, the gradient reversal layer guided domain adversarial mechanism effectively filters out tissue background interference, and the migration learning strategy realizes rapid adaptation of the model in a complex physiological environment, thereby providing a full-view, non-contact and high-precision new method for blood flow monitoring.
Need to check novelty before this filing date? Find Prior Art