CN-121999293-A - Intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage

CN121999293ACN 121999293 ACN121999293 ACN 121999293ACN-121999293-A

Abstract

The invention relates to the technical field of medical image processing and intelligent medical monitoring, in particular to an intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage. The system comprises a time sequence image management module, a dynamic change quantization module and an intelligent early warning judgment module, wherein the time sequence image management module takes a leakage point image coordinate P0 in reference input as a space anchor point, performs non-rigid registration on a multi-time point follow-up image and the reference image, outputs a time sequence image sequence with time-space alignment, generates a multi-dimensional quantization parameter set by fusing and calculating morphology, density and texture characteristics based on the sequence and a reference segmentation mask M0, and outputs a grading early warning signal by matching the parameter set with a preset early warning rule base comprising multi-parameter coupling logic. The invention realizes the technical span from static analysis to dynamic monitoring of cerebrospinal fluid leakage, and improves the accuracy, objectivity and efficiency of state sensing through multi-parameter fusion quantization and intelligent judgment.

Inventors

HE XIAOFENG
ZHANG XIAOBO
ZHANG XIN
WEI YINGTIAN
ZHANG ZHONGLIANG
ZHANG XIAO
XIAO YUEYONG
WANG SIBIN
GUO XINYUAN

Assignees

中国人民解放军总医院第一医学中心

Dates

Publication Date: 20260508
Application Date: 20260129

Claims (10)

1. An intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage, which is characterized by comprising: the time sequence image management module is configured to take the coordinates P0 of the leakage point images in the reference input as space anchor points, perform non-rigid registration on the follow-up images acquired at different time points and the reference images, and output time sequence image sequences aligned in time and space; the dynamic change quantization module is configured to generate a multidimensional quantization parameter set for representing dynamic change by fusing and calculating morphology, density and texture characteristics based on the time sequence image sequence and a drain region segmentation mask M0 in the reference input; And the intelligent early warning judgment module is configured to match the multi-dimensional quantitative parameter set with a preset early warning rule base, wherein the early warning rule base comprises rules defined based on logic combination of a plurality of parameters in the multi-dimensional quantitative parameter set, and outputs corresponding grading early warning signals.
2. The intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage according to claim 1, wherein the specific way for the time sequence image management module to perform registration is as follows: defining a three-dimensional region of interest for driving registration optimization by taking the leakage point image coordinate P0 as a center; And in the three-dimensional interested region, solving a nonlinear space transformation field $T_ { 0\ rightarrow T } $ for mapping the follow-up image It to the space of the reference image I0 by adopting symmetric normalization transformation based on mutual information maximization.
3. The intelligent monitoring and early warning system for dynamic changes of cerebrospinal fluid leakage according to claim 2, wherein said time-series image management module is further configured to: after the follow-up image It is acquired, the initial similarity between the follow-up image It and the reference image I0 in the three-dimensional interested area is automatically checked, if the initial similarity is lower than a preset threshold value, a multi-resolution searching strategy of a registration algorithm is preferentially adjusted, and then the nonlinear space transformation field is solved.
4. The intelligent monitoring and early warning system for dynamic changes of cerebrospinal fluid leakage according to claim 1, wherein said multi-dimensional quantization parameter set at least comprises a volume change rate VCR, an average gray value change amount $\ DELTA MEAN $ and a texture entropy value change amount $\ Delta Entropy $.
5. The intelligent monitoring and early warning system for dynamic changes in cerebrospinal fluid leakage according to claim 4, wherein said volume change rate VCR is calculated by: on the registered image $I_t' $, the coordinate P0 is taken as a search center, and a drain region segmentation mask Mt of a current time point is generated by calling a segmentation model based on nnU-Net architecture; The rate of change of volume is calculated according to the formula $vcr= \frac { v_t-v_0} { v_0} \times 100% $, where $v_0$ and $v_t$ are the physical volumes represented by the masks $m_0$ and $m_t$, respectively.
6. The intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage according to claim 4, wherein the calculation process of the texture entropy change amount $\ Delta Entropy $ comprises: within the common area defined by the masks $M_0$and $M_t$, firstly, carrying out gray scale normalization processing on the image blocks to eliminate intensity differences, and then calculating a gray scale co-occurrence matrix of the image blocks; According to the formula of $ Entropy = - \sum_ { i=1 } { n_g } \sum_ { j=1 } { n_g } p (i, j) \log_ p (i, j) $ the entropy value of the normalized gray level co-occurrence matrix is calculated, and the entropy value variation $\ Delta Entropy $ is obtained.
7. The intelligent monitoring and early warning system for dynamic changes of cerebrospinal fluid leakage according to claim 1, wherein the early warning rule base in the intelligent early warning decision module is dynamically updated, configured to receive feedback information of the early warning result from the user, and based on the feedback, adaptively adjust the threshold value or the logic combination in the rule by a reinforcement learning algorithm.
8. The intelligent monitoring and early warning system for dynamic changes of cerebrospinal fluid leakage according to claim 4, wherein said early warning rule base comprises one or more of the following rules: if the volume change rate VCR is more than twenty-five percent, the early warning level is "significant progress"; if the volume change rate VCR is more than fifteen percent and the texture entropy change amount $\ Delta Entropy $ is more than zero, the early warning level is 'progress attention'; If the volume change rate VCR is less than minus twenty percent and the absolute value of the average gray value change amount $\ DELTA MEAN $ is greater than the preset sensitivity threshold, the early warning level is "significant absorption".
9. The intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid according to any one of claims 1 to 8, further comprising a data quality evaluation module configured to evaluate the signal-to-noise ratio and contrast ratio of the registered image $i_t' $in the vicinity of the reference coordinate P0 before the calculation by the dynamic change quantification module, and trigger data retrieval or prompt manual review if the evaluation result does not meet the preset quality standard.
10. A computer readable storage medium having stored thereon a computer program, wherein the program when executed by a processor performs the function of an intelligent monitoring and early warning system for dynamic changes in cerebrospinal fluid according to any one of claims 1 to 9.

Description

Intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage Technical Field The invention relates to the technical field of medical image processing and intelligent medical monitoring, in particular to an intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage. Background In the prior art, methods and systems exist that enable intelligent identification and initial positioning of cerebrospinal fluid leakage points. For example, through the cooperation of multi-mode image registration (such as ANTs algorithm) and artificial intelligent models (such as nnUnet and Yolov 8), the accurate segmentation of the cerebrospinal fluid leakage area in the single-time-point image and the positioning of the leakage point coordinates are realized, and the problems of identification and positioning of the initial state are effectively solved. However, imaging of the leakage of cerebrospinal fluid is a dynamic process. The above prior art and other similar schemes focus on static, disposable analysis, and cannot meet the technical requirements of long-term, dynamic, quantitative tracking of a target area. Specifically, the prior art has the following technical drawbacks: The prior scheme is an isolated analysis on single-time acquired images, and the system architecture cannot automatically correlate and compare the sequential image data of the same target at a plurality of time points, so that continuous data describing the evolution of the target cannot be generated. The change evaluation relies on manual comparison, namely, for images at different time points, an operator is required to compare and analyze the images in a manual mode so as to qualitatively judge the change of a target area. The process has low efficiency, and the interference of the dominant factors is large, so that the consistency of analysis results is poor and the reliability is low. No objective quantitative early warning mechanism is provided, and the prior art cannot establish an automatic state evaluation and prompt mechanism based on data because the system cannot automatically extract and calculate quantitative evolution parameters (such as volume change rate and image characteristic value change) of a target area. It is difficult to immediately recognize a significant change in state from the technical level. The state assessment lacks continuous data support, and the system cannot provide objective and quantitative data chains based on time sequence images for the overall state of the target area. Operators have difficulty in accurately grasping dynamic details, and can not provide efficient and reliable data support for subsequent comprehensive analysis. Disclosure of Invention In view of the above, the invention provides an intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage, which aims to construct an analysis model for quantitatively tracking the dynamic change of the dynamic change and executing intelligent early warning by taking the coordinates of a leakage point image and a segmentation area determined by first intelligent identification and positioning as reference input, and finally realize intelligent monitoring and early warning for the dynamic change of cerebrospinal fluid leakage. In order to achieve the above purpose, the present invention adopts the following technical scheme: In a first aspect, the present invention provides an intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage, comprising: In a specific embodiment of the present invention, In a specific embodiment of the present invention, In a second aspect, the present invention provides an intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage, which is applied to the above-mentioned In a third aspect, the invention provides an intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage, Compared with the prior art, the intelligent monitoring and early warning system for dynamic change of cerebrospinal fluid leakage breaks through the technical paradigm that the prior art is limited to isolated analysis of single-time-point images, and initially constructs a dynamic monitoring and early warning system for a target area, and has the following beneficial effects: 1. system capability crossover from "static analysis" to "dynamic monitoring The prior art provides a static snapshot of the target at a certain moment, and the dynamic evolution map of the target in a time dimension is generated by systematically introducing time sequence image analysis. The transition enables the system to be upgraded from providing the 'space state' information to providing the 'space-time evolution' information, provides unprecedented continuous data support for comprehensively grasping the dynamic process of the target object, and realizes the f