Liver and gall system interactive three-dimensional visual reconstruction system based on volume rendering technology

Abstract

The invention discloses a volume rendering technology-based interactive three-dimensional visualization reconstruction system for a liver and gall system, which belongs to the technical field of medical image three-dimensional visualization and comprises an organ perception optimization transfer function adaptive optimization module, a multi-level adaptive ray casting rendering module, a virtual anatomy interaction control module and a quantitative measurement labeling module. An organ perception optimization transmission function adaptive optimization module performs probability modeling on CT value distribution of different tissues of the liver and gall system based on a Gaussian mixture model, and automatically generates an optimization transmission function; the multi-level self-adaptive ray casting rendering module adopts GPU acceleration and a self-adaptive sampling strategy to realize efficient volume rendering; the virtual anatomy interaction control module supports plane cutting and transparency adjusting operation; and the quantitative measurement labeling module provides distance, angle and volume measurement functions, and triggers adaptive adjustment of an optimized transmission function through a closed-loop feedback mechanism, so that the display quality and interaction performance of three-dimensional visualization of the liver and gall system are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical image three-dimensional visualization, in particular to an interactive three-dimensional visualization reconstruction system of hepatobiliary system based on volume rendering technology. BACKGROUND

[0002] Three-dimensional visualization reconstruction technology of hepatobiliary system has important application value in clinical diagnosis, surgical planning and medical education. With the development of computer tomography technology, three-dimensional reconstruction of hepatobiliary system based on CT image data has become an important supporting technology for precision surgery.

[0003] Chinese patent CN119478203A discloses a reconstruction method of hepatobiliary pancreatic three-dimensional image. The method introduces a U-Net variant segmentation network for liver, gallbladder and pancreas, uses a convolutional neural network to segment multi-modal images, adopts a deep learning driven non-rigid registration algorithm for spatial alignment, and realizes three-dimensional model construction through adaptive three-dimensional grid reconstruction and texture mapping. However, the prior art has the following shortcomings:

[0004] In terms of visualization method, the prior art uses an indirect volume rendering method based on surface reconstruction, which needs to convert volume data into a mesh model before rendering. This indirect method has obvious limitations in dealing with the complex internal structure of the hepatobiliary system. Specifically, the CT value distribution of liver parenchyma, gallbladder wall, bile duct and blood vessel system is obviously different but adjacent to each other. The mesh reconstruction method is difficult to clearly present these interwoven anatomical structures at the same time, which easily causes detail loss and structure occlusion.

[0005] In terms of interactive control, the visualization result of the prior art is a static three-dimensional mesh model, and users cannot observe the internal structure of the liver in real time by adjusting the transparency and cutting plane, which makes it difficult to meet the needs of virtual dissection exploration and pathological analysis. Especially in the preoperative discussion of complex hepatobiliary surgery, surgeons need to observe the running and spatial relationship of the intrahepatic duct system from multiple angles and levels, and the static mesh model cannot provide such interactive exploration capability.

[0006] In terms of rendering efficiency, the mesh generation and optimization process of the prior art has large amount of calculation, especially when dealing with high-resolution CT data, which needs to go through multiple time-consuming steps such as boundary extraction, point cloud generation, triangulation, mesh optimization, etc. This offline processing mode cannot realize real-time visualization parameter adjustment, which limits the interactive response speed in clinical application.

[0007] At the level of optimizing the transfer function design, the prior art lacks adaptive optimization mechanism for different tissue characteristics of the hepatobiliary system. There are specific rules in the CT value distribution of liver parenchyma, bile duct, blood vessels and lesion area. The prior art fails to fully utilize the prior knowledge to automatically optimize the visualization parameters, and a large amount of manual adjustment is required to obtain satisfactory display effect.

[0008] At the level of quantitative analysis, the prior art mainly focuses on the geometric reconstruction of the three-dimensional model, lacks integrated measurement marking tool support, and is difficult to directly perform distance measurement, angle calculation and volume marking in the visualization environment, which limits its practical value in clinical diagnosis and teaching training.

[0009] In summary, the existing three-dimensional reconstruction technology of the hepatobiliary system has obvious deficiencies in visualization method, interaction ability, rendering efficiency, parameter optimization and quantitative analysis, and it is urgent to develop an interactive three-dimensional visualization reconstruction system based on volume rendering technology to meet the needs of digital teaching of hepatobiliary medicine and development of precise surgery. SUMMARY

[0010] In view of the problems existing in the prior art, the present application provides an interactive three-dimensional visualization reconstruction system of the hepatobiliary system based on volume rendering technology, which realizes high-quality three-dimensional visualization and real-time interactive operation of the hepatobiliary system through the deep coupling and collaborative architecture of four core modules of organ perception optimization transfer function adaptive optimization module, multi-level adaptive ray casting rendering module, virtual dissection interactive control module and quantitative measurement marking module.

[0011] The technical scheme adopted by the present application is an interactive three-dimensional visualization reconstruction system of the hepatobiliary system based on volume rendering technology, which comprises an organ perception transfer function adaptive optimization module, a multi-level adaptive ray casting rendering module, a virtual dissection interactive control module and a quantitative measurement marking module.

[0012] The organ perception transfer function adaptive optimization module is used for receiving CT body data of the hepatobiliary system, probabilistically modeling the CT value distribution of liver parenchyma, gallbladder, bile duct and blood vessel structure based on Gaussian mixture model, generating initial transfer function parameters according to the tissue-specific probability distribution characteristics, and outputting the optimized transfer function containing transparency mapping and color mapping to the multi-level adaptive ray casting rendering module.

[0013] The multi-level adaptive ray casting rendering module is used for receiving the optimized transfer function and the CT body data, performing volume rendering rendering by using GPU accelerated ray casting algorithm, dynamically adjusting the sampling density according to the viewpoint distance and local gradient change, combining the multi-level detail technology to realize smooth interactive response while ensuring high-quality display effect, and outputting the rendering result to the display terminal.

[0014] The virtual anatomy interactive control module is used to respond to the user's interactive operation commands. It realizes real-time visualization control of the internal structure of the liver by defining cutting plane parameters and transparency adjustment parameters, and feeds the cutting plane parameters and transparency adjustment parameters back to the multi-level adaptive ray projection rendering module to update the rendering results.

[0015] The quantization measurement and annotation module is used to perform distance measurement, angle calculation, and volume annotation operations in a 3D visualization scene, and to overlay the measurement results onto the rendering results.

[0016] The invention also includes a closed-loop feedback optimization mechanism, in which the measurement results of the quantitative measurement and annotation module are transmitted back to the organ perception transfer function adaptive optimization module to evaluate the current visualization effect and trigger the adaptive adjustment of the optimized transfer function.

[0017] The beneficial effects of this invention include:

[0018] By employing direct volume rendering technology to replace traditional surface reconstruction methods, this invention can simultaneously and clearly present the parenchymal tissues, ductal structures, and vascular networks of the hepatobiliary system, avoiding the loss of details during mesh reconstruction and significantly improving the visualization quality of complex anatomical structures.

[0019] By constructing an adaptive optimization mechanism for the optimized transfer function of organ perception, this invention can automatically adjust transparency and color mapping according to the CT value distribution characteristics of different tissues in the hepatobiliary system, reducing the workload of manual parameter adjustment and enabling users to quickly obtain the best visualization effect.

[0020] By employing a rendering strategy that combines GPU-accelerated ray casting algorithms with multi-level detail technology, this invention can achieve smooth interactive response while ensuring high-quality display effects, and supports real-time viewpoint changes and parameter adjustments.

[0021] By integrating virtual anatomy functionality, this invention enables users to observe the internal structure of the liver in real time by adjusting the cutting plane and transparency, providing an intuitive and interactive environment for virtual anatomical exploration and pathological analysis.

[0022] By integrating measurement and annotation tools and establishing a closed-loop feedback optimization mechanism, this invention combines visualization effect evaluation with parameter optimization to form a complete closed loop of adaptive adjustment, further improving the intelligence level of the visualization system. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall structure of the system of the present invention, wherein label 1 represents the organ perception optimization transfer function adaptive optimization module, label 2 represents the multi-level adaptive ray projection rendering module, label 3 represents the virtual anatomy interactive control module, and label 4 represents the quantitative measurement annotation module.

[0024] Figure 2 This is a schematic diagram of the internal structure of the organ perception optimization transfer function adaptive optimization module.

[0025] Figure 3 This is a schematic diagram of the workflow of the multi-layered adaptive ray casting rendering module.

[0026] Figure 4 This is a functional structure diagram of the virtual anatomy interactive control module.

[0027] Figure 5 This is a functional architecture diagram of the quantitative measurement and labeling module. Detailed Implementation

[0028] Please refer to the attached document. Figures 1-5 The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0029] like Figure 1 As shown, the interactive 3D visualization reconstruction system for the hepatobiliary system based on volume rendering technology provided by this invention includes an organ-sensing optimized transfer function adaptive optimization module 1, a multi-level adaptive ray casting rendering module 2, a virtual anatomy interactive control module 3, and a quantitative measurement and annotation module 4. These four modules establish deep coupling relationships at both the parameter and state levels, forming a closed-loop collaborative system that combines forward transmission and reverse feedback.

[0030] The organ-sensing optimized transfer function adaptive optimization module 1 is located at the data processing front end of the system, receiving CT volume data of the hepatobiliary system from a CT scanning device or a DICOM data interface as input. In one embodiment of the invention, the design of this module fully considers the differences in CT value distribution among different tissues of the hepatobiliary system. Figure 2 As shown, this module mainly consists of four parts: a tissue CT value statistical analysis unit, a probability distribution modeling unit, an optimized transfer function parameter generation unit, and an adaptive adjustment unit.

[0031] The tissue CT value statistical analysis unit is responsible for performing voxel-level statistical analysis on the input CT volume data, extracting the CT value distribution characteristics of various tissue structures in the hepatobiliary system. In one possible implementation, this unit first preprocesses the volume data, including noise filtering and window width and level adjustment, then calculates the overall CT value histogram and identifies the main peak regions. The typical CT value range for liver parenchyma is approximately 40-60 HU, the CT value of gallbladder contents is close to that of water (approximately 0-20 HU), the CT value of bile duct walls and blood vessel walls is approximately 30-50 HU, while the enhanced CT value of intrahepatic vessels can reach 100-150 HU. By analyzing these characteristic CT value distributions, the tissue CT value statistical analysis unit provides a data foundation for subsequent probabilistic modeling.

[0032] The probability distribution modeling unit performs probabilistic modeling of the CT value distribution of different tissues in the hepatobiliary system based on Gaussian mixture models. This invention proposes an organ-specific Gaussian mixture model construction method. This method, tailored to the anatomical characteristics of the hepatobiliary system, employs a four-component Gaussian mixture model corresponding to four main tissue structures: liver parenchyma, gallbladder, bile ducts, and blood vessels. In one embodiment of this invention, the Gaussian mixture model parameters of the probability distribution modeling unit are iteratively optimized using an expectation-maximization algorithm to ultimately obtain the mean, variance, and mixture weight parameters for each tissue category.

[0033] The organ-specific Gaussian mixture model proposed in this invention can be implemented as follows: For any voxel in the volume data, the posterior probability that its CT value belongs to the k-th tissue class is calculated using Bayes' theorem. Specifically, given a CT value x, the probability that it belongs to the k-th tissue class is obtained by dividing the product of the prior probability and the conditional probability density of that tissue class by the total probability. This invention sets up four Gaussian components for the hepatobiliary system, corresponding to the liver parenchyma, gallbladder, bile duct, and vascular structures, respectively. Each Gaussian component is determined by three parameters: mean, variance, and mixture weight.

[0034] To accurately describe the CT value distribution characteristics of hepatobiliary system tissues, this invention proposes a tissue-specific Gaussian mixture model, whose probability density function is defined as:

[0035] ,

[0036] in, The CT value of the voxel. The total number of organizational categories in the hepatobiliary system visualization scenario. These correspond to four types of tissues: liver parenchyma, gallbladder, bile ducts, and blood vessels. For the first The mixed weighting coefficients of the class of organizations satisfy the following conditions: and , For the first Mean parameters of the CT value distribution of tissue-like structures. For the first The variance parameter of the CT value distribution of tissue-like organisms. The mean is variance is The Gaussian probability density function.

[0037] In one embodiment of the present invention, the mean of the Gaussian component corresponding to the liver parenchyma. Set to 50HU, variance Set to 100HU², mixed weights Setting this parameter to 0.65 reflects the anatomical characteristic that the liver parenchyma occupies the majority of the volume in the hepatobiliary system. The mean of the Gaussian component corresponding to the gallbladder. Set to 10 HU, variance Set to 64HU², mixed weights Set to 0.10. Mean Gaussian component of the bile duct. Set to 40 HU, variance Set to 81HU², mixed weights Set to 0.08. Mean Gaussian component corresponding to blood vessels. Set to 120 HU (under enhanced CT conditions), variance Set to 225HU², mixed weights Set it to 0.17.

[0038] The optimized transfer function parameter generation unit generates initial optimized transfer function parameters based on probability distribution modeling results. The optimized transfer function is a core component of volume rendering technology, mapping the CT values ​​of voxels to transparency and color attributes. This invention proposes an automatic optimized transfer function generation method based on tissue posterior probability. This method automatically determines the transparency mapping curve and color mapping scheme according to the posterior probability distribution of each tissue category.

[0039] The transparency mapping function proposed in this invention is defined as follows:

[0040] ,

[0041] in, CT value The transparency value of the voxel, with a range of values ​​of [value range missing]. , The optimal value for the maximum transparency parameter is 0.95. This is a transparency adjustment factor used to control the sensitivity to changes in transparency; the preferred value range is 1.5-3.0. CT value Belongs to the The posterior probability of a class of organizations, This indicates taking the largest posterior probability value among all organizational categories.

[0042] The design principle of the above transparency mapping function is that when the CT value of a voxel clearly belongs to a certain type of tissue, its maximum posterior probability is close to 1, and at this time the transparency value approaches 1. This makes the voxel clearly visible in the rendering result; when the CT value of the voxel is in the boundary area of ​​multiple tissues, its maximum a posteriori probability is low, and the transparency value is reduced accordingly, so that the boundary area presents a natural transition effect.

[0043] The color mapping scheme employs a weighted mixing strategy based on tissue category. This invention assigns characteristic colors to four types of tissues in the hepatobiliary system: the liver parenchyma is painted a deep reddish-brown to match its anatomical color; the gallbladder is painted a pale yellowish-green to highlight its cystic nature; the bile ducts are painted a bright green to enhance recognizability; and blood vessels are painted a bright red to conform to medical visualization conventions.

[0044] The color mapping function proposed in this invention is defined as follows:

[0045] ,

[0046] in, CT value The mapped color values ​​of the voxels are represented as RGB three-channel vectors. CT value Belongs to the The posterior probability of a class of organizations, For the first The characteristic color vector of the organization, This represents the total number of organization categories.

[0047] In one embodiment of the present invention, the characteristic color of liver parenchyma Set to RGB (139, 69, 49), the characteristic color of the gallbladder. Set to RGB (189, 183, 107), the characteristic color of the bile duct. Set to RGB (60, 179, 113), the characteristic color of blood vessels. Set to RGB (220, 20, 60). Through posterior probability weighted mixing, voxels located in tissue boundary regions can achieve a smooth color mapping effect.

[0048] The adaptive adjustment unit receives feedback information from the quantization measurement and annotation module 4 and dynamically adjusts and optimizes the transfer function parameters based on the visualization effect evaluation results. In the closed-loop feedback optimization mechanism, when the quantization measurement and annotation module 4 detects insufficient visibility of the target tissue or unclear structural boundaries, the adaptive adjustment unit automatically adjusts the transparency adjustment factor for the corresponding tissue category. and mixed weights To optimize the visualization.

[0049] The multi-layered adaptive ray casting rendering module 2 is the core module of this invention for achieving high-quality volume rendering. For example... Figure 3 As shown, this module employs a GPU-accelerated ray casting algorithm, combined with multi-level detail technology, to achieve efficient volume rendering. In one embodiment of the invention, the module mainly consists of five parts: a ray generation unit, an adaptive sampling unit, a voxel interpolation unit, a color accumulation unit, and a multi-level detail control unit.

[0050] The ray generation unit is responsible for generating sampled rays that traverse the volume data space based on the current viewpoint position and viewing direction. In one possible implementation, the ray generation unit employs a screen-space-based ray emission strategy to generate a sampled ray for each pixel of the rendered image, originating from the viewpoint and passing through that pixel. The starting point of the ray is determined by the viewpoint position, and the direction of the ray is jointly determined by the viewpoint position, pixel position, and projection parameters.

[0051] The adaptive sampling unit is one of the innovative technical points of this invention, responsible for dynamically adjusting the sampling density based on local gradient changes and viewpoint distance. Traditional ray casting algorithms use a fixed step size for sampling, which generates excessive redundant sampling in uniform regions and may result in insufficient sampling in regions rich in detail. This invention proposes a gradient-aware adaptive sampling strategy that can intelligently adjust the sampling step size according to the local structural complexity.

[0052] The adaptive sampling step size calculation method proposed in this invention is defined as follows:

[0053] ,

[0054] in, For position Adaptive sampling step size at the location, The optimal value for the base sampling step size is 0.5 times the diagonal length of the voxel. This is the gradient sensitivity coefficient, used to control the degree of influence of the gradient on the sampling step size. The preferred value range is 0.5-2.0. For position The gradient magnitude of the volume data at that location. This is the distance attenuation coefficient, used to adjust the sampling density according to the viewpoint distance. The preferred value range is 0.3-0.8. For position Euclidean distance to the viewpoint The diagonal length of the volume data space is used to normalize the distance parameter.

[0055] The physical meaning of the above adaptive sampling step size formula is that when the local gradient magnitude is large (indicating the presence of tissue boundaries or structural details), the sampling step size is reduced accordingly to capture more detailed information; when the distance from the viewpoint is far, the sampling step size is appropriately increased to reduce unnecessary computational overhead. This strategy is consistent with the visual characteristics of the human eye, which has reduced sensitivity to details of distant objects.

[0056] The voxel interpolation unit is responsible for obtaining the corresponding CT value from the volume data based on the sampling point location. Since the sampling point usually does not precisely coincide with the voxel center, interpolation calculation is required to obtain the CT value at the sampling point. In one embodiment of the present invention, the voxel interpolation unit employs a trilinear interpolation method, which achieves a good balance between accuracy and efficiency. Trilinear interpolation first performs linear interpolation in the x-direction to obtain four intermediate values, then interpolates in the y-direction to obtain two intermediate values, and finally interpolates in the z-direction to obtain the final sampled value.

[0057] The color accumulation unit is responsible for accumulating the color and transparency contributions of each sampling point along the light direction to generate the final pixel color value. This invention adopts a forward accumulation method, starting from the entry point of the light entering the volume data, processing the color contribution of each sampling point sequentially and accumulating it into the final result.

[0058] The forward color accumulation formula used in this invention is defined as follows:

[0059] ,

[0060] ,

[0061] in, The accumulated color value. The currently accumulated color values, This is the accumulated transparency value. This represents the currently accumulated transparency value. For the first The color value of each sample point is determined by the color mapping of the optimized transfer function. For the first The transparency value of each sampling point is determined by the transparency mapping of the optimized transfer function.

[0062] The process of color accumulation from , Initially, each sampling point is processed sequentially along the direction of the light. When... When the value is close to 1 (the preferred threshold is 0.99), it means that the light has been fully absorbed and the contribution of subsequent sampling points can be ignored. At this point, the accumulation process can be terminated in advance to improve rendering efficiency.

[0063] The multi-level detail control unit is responsible for dynamically switching the rendering precision level based on viewpoint distance and interaction status. This invention employs a three-level multi-level detail strategy, corresponding to high-precision mode, medium-precision mode, and low-precision mode. High-precision mode uses volume data at the original resolution and the minimum sampling step size, suitable for detailed observation in a static state; medium-precision mode uses volume data downsampled by 2 times and a sampling step size of 1.5 times, suitable for real-time response during slow interactions; low-precision mode uses volume data downsampled by 4 times and a sampling step size of 2 times, suitable for smooth display during rapid rotation and scaling operations.

[0064] The multi-level detail switching strategy proposed in this invention is based on two metrics: interaction speed and frame rate. When a user is detected interacting and the interaction speed exceeds a set threshold, the system automatically switches to a lower precision mode to ensure smooth response. When the interaction stops, the system automatically reverts to a high precision mode after a short delay to present the best display effect. This strategy achieves a dynamic balance between display quality and interaction response.

[0065] The virtual anatomy interactive control module 3 is a key module for achieving interactive visualization in this invention. For example... Figure 4 As shown, this module provides two core functions: defining the cutting plane and interactively adjusting transparency, enabling users to control the display effect of the three-dimensional visualization of the hepatobiliary system in real time. In one embodiment of the invention, the module mainly consists of three parts: a cutting plane control unit, a transparency adjustment unit, and an interactive event processing unit.

[0066] The cutting plane control unit responds to the user's defined cutting plane operation, supporting planar cutting in any direction and position. In one possible implementation, the cutting plane is determined by two parameters: a plane normal vector and a point on the plane. The user can adjust the position and orientation of the cutting plane by dragging with the mouse, and the system updates the cutting results in real time and displays the tissue cross-section on the cutting plane. The cutting plane is implemented using a geometric trimming method, setting the area outside the cutting plane to be completely transparent during light projection, thereby achieving a virtual anatomical effect.

[0067] This invention supports the combination of multiple cutting planes, allowing users to define up to six cutting planes simultaneously to form cropping regions of arbitrary shapes. The combination of multiple cutting planes uses a logical AND operation; only regions located on the valid sides of all cutting planes are rendered and displayed. This feature enables users to precisely define the anatomical region of interest, eliminating occlusion from surrounding interfering structures.

[0068] The transparency adjustment unit provides independent transparency control for each tissue type. In one embodiment of the invention, the user can adjust the transparency parameters of four tissue types—liver parenchyma, gallbladder, bile duct, and blood vessels—using a slider. The transparency adjustment employs a multiplicative correction method, multiplying the user-set transparency correction coefficient by the original transparency value output by the optimized transfer function to obtain the final display transparency.

[0069] The transparency correction calculation method proposed in this invention is defined as follows:

[0070] ,

[0071] in, This is the corrected transparency value. To optimize the raw transparency value of the transfer function output, For the first The user-defined transparency correction factor for the organization type has a range of values. ,when At that time, such organizations were completely transparent and invisible. Maintain the original transparency settings.

[0072] The transparency adjustment unit also provides a quick global transparency adjustment function, allowing users to simultaneously reduce the transparency of all tissues with a single operation, making internal structures visible through the outer layers of tissue. This function is particularly useful for observing the course and distribution of the intrahepatic duct system, helping users understand complex spatial anatomical relationships.

[0073] The interaction event handling unit is responsible for capturing and parsing user interactions, including mouse events, keyboard events, and touch events, and converting the parsed results into cutting plane parameters and transparency adjustment parameters. This unit is also responsible for maintaining the interaction state machine, tracking the user's current operation mode (such as rotation mode, translation mode, cutting mode, measurement mode, etc.), and passing state information to other modules to coordinate system behavior.

[0074] The quantitative measurement and labeling module 4 is the core module of this invention that provides accurate quantitative analysis capabilities. For example... Figure 5 As shown, this module integrates three main functions: distance measurement, angle calculation, and volume annotation, providing precise quantitative analysis support for clinical diagnosis and teaching training. In one embodiment of the invention, the module mainly consists of four parts: a distance measurement unit, an angle calculation unit, a volume annotation unit, and a result visualization unit.

[0075] The distance measurement unit provides distance measurement functionality between any two points in three-dimensional space. In one possible implementation, the user selects a start and end point by clicking, and the system automatically calculates the Euclidean distance between the two points and displays the measurement result in millimeters. The distance measurement unit supports the simultaneous display of multiple sets of measurement data, with each measurement result distinguished by connecting lines of different colors and labeled text. To improve measurement accuracy, this unit employs a sub-voxel-level positioning method, automatically locking onto the nearest tissue surface point through gradient search near the clicked location.

[0076] The angle calculation unit provides the function of measuring the angle formed by any three points in three-dimensional space. Users select three positioning points sequentially, with the second point as the vertex of the angle. The system automatically calculates the angle formed by the vector from the first point to the second point and the vector from the second point to the third point, and displays the measurement result in degrees. This function is particularly suitable for measuring anatomical parameters such as the bifurcation angle of blood vessels and the confluence angle of bile ducts.

[0077] The volume annotation unit provides volume measurement and annotation functions for regions of interest. This invention supports two volume measurement methods: one is automatic volume calculation based on threshold segmentation, where the system automatically counts the total number of voxels that meet the criteria and converts them into volume values ​​after the user specifies a CT value threshold range; the other is interactive volume calculation based on region growing, where the system automatically grows connected regions of the same type of tissue and calculates their volume after the user clicks on a seed point. The volume calculation results are displayed in cubic centimeters or milliliters and can be annotated at corresponding locations in the 3D visualization scene.

[0078] The results visualization unit is responsible for overlaying measurement results onto the volume rendering image. This unit employs a layered rendering strategy, overlaying 2D graphic elements such as measurement line segments, angle markers, volume boundaries, and numerical annotations after the volume rendering is complete. To ensure the visibility of measurement annotations in complex 3D scenes, this unit uses depth testing and occlusion detection technology. When annotations are occluded by structures, the unit automatically adjusts the annotation position or displays the occluded portion as a dashed line.

[0079] This invention establishes a complete closed-loop feedback optimization mechanism to achieve adaptive adjustment of the visualization effect. In this mechanism, the measurement results of the quantification measurement annotation module 4 are transmitted back to the organ perception transfer function adaptive optimization module 1 to evaluate the current visualization effect and trigger adaptive adjustment of the optimized transfer function.

[0080] The workflow of the closed-loop feedback optimization mechanism is as follows: When performing measurement operations, the quantitative measurement and annotation module 4 simultaneously evaluates the visualization quality indicators of the target tissue region, including parameters such as tissue boundary clarity, contrast, and visibility. When the evaluation result is lower than a preset quality threshold, a feedback signal is transmitted to the adaptive adjustment unit of the organ perception transfer function adaptive optimization module 1. The adaptive adjustment unit determines the tissue category requiring enhanced display based on the feedback information and automatically adjusts the corresponding optimized transfer function parameters, such as increasing the transparency adjustment factor of that type of tissue or adjusting its blending weight. The optimized transfer function with adjusted parameters is then transmitted to the multi-layer adaptive ray casting rendering module 2 to generate an updated visualization result.

[0081] In one embodiment of the present invention, the assessment of tissue boundary sharpness employs a gradient magnitude statistical method, calculating the average gradient magnitude of the target tissue boundary region as a sharpness index. When this index is below 0.3 (normalized value), the boundary is determined to be unclear, triggering an adaptive adjustment of the optimized transfer function, increasing the transparency adjustment factor of the corresponding tissue by 15% to 30% to enhance boundary contrast.

[0082] The coupling relationship and synergistic effect among the four core modules of this invention are reflected in the following aspects:

[0083] At the parameter-level coupling level, the optimized transfer function parameters output by the organ-aware transfer function adaptive optimization module 1 directly serve as key input parameters for the multi-layered adaptive ray casting rendering module 2, establishing a tight data dependency between the two modules. Similarly, the cutting plane parameters and transparency adjustment parameters output by the virtual anatomy interactive control module 3 also serve as input parameters for the rendering module, enabling real-time influence of interactive control on the rendering results.

[0084] At the state-level coupling level, the interaction state information maintained by the interaction event processing unit is shared among multiple modules, enabling each module to adjust its behavior according to the current interaction mode. When the user is in measurement mode, the multi-layer adaptive ray casting rendering module 2 automatically switches to high-precision mode to ensure measurement accuracy; when the user performs rapid interactive operations, this module automatically reduces the precision to ensure response speed.

[0085] At the closed-loop feedback level, the evaluation results of the visualization effect by the quantitative measurement and annotation module 4 are transmitted back to the organ perception transfer function adaptive optimization module 1, forming a complete closed loop that combines forward transmission and reverse feedback, thereby achieving adaptive optimization of the visualization parameters.

[0086] At the synergistic level, the functions of the four modules promote each other, creating a synergistic effect. The organ perception optimization transfer function provides accurate tissue classification, laying the foundation for the transparency layering control of virtual anatomy; the adaptive sampling strategy ensures rendering quality while providing smooth response for interactive operations; the virtual anatomy function enables quantitative measurements to accurately annotate specific anatomical layers; and the feedback information from quantitative measurements further adjusts and optimizes the transfer function parameters, forming a virtuous cycle of continuous improvement.

[0087] The technical solution of this invention demonstrates significant technical effects in practical applications. Taking common clinical cases of hepatobiliary stones as an example, this invention can clearly display the location and size of the stones in the bile duct system, observe the spatial relationship between the stones and surrounding tissues through virtual anatomy, and accurately calculate the stone volume and the degree of bile duct dilation using measurement tools, providing accurate anatomical information for surgical planning. Compared with existing mesh-based reconstruction methods, the volume rendering technology of this invention can simultaneously display the stones and the duct wall structure within the bile duct lumen, avoiding the limitation of mesh models in representing semi-transparent internal structures.

[0088] The rendering efficiency tests of this invention show that, on a computer equipped with a mid-range GPU (such as an NVIDIA RTX 3060), the system can achieve an interactive frame rate of over 30fps for CT volume data at a resolution of 512×512×256, meeting the requirements for smooth interactive operation. In high-precision static display mode, the rendering quality is comparable to that of a professional medical imaging workstation, with clear tissue boundaries and natural color transitions, meeting the quality requirements for clinical diagnosis and medical education.

[0089] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. An interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology, characterized in that, include: The organ-sensing optimized transfer function adaptive optimization module is used to receive CT data of the hepatobiliary system, perform probabilistic modeling of the CT value distribution of liver parenchyma, gallbladder, bile duct and vascular structures based on Gaussian mixture model, generate initial optimized transfer function parameters according to tissue-specific probability distribution characteristics, and output the optimized transfer function including transparency mapping and color mapping. The multi-layer adaptive ray casting rendering module is used to receive the optimized transfer function and the CT volume data, perform volume rendering using a GPU-accelerated ray casting algorithm, dynamically adjust the sampling density according to the viewpoint distance and local gradient changes, and generate rendering results by combining multi-layer detail technology and output them to the display terminal. The virtual anatomy interactive control module is used to respond to the user's interactive operation commands. It realizes real-time visualization control of the internal structure of the liver by defining cutting plane parameters and transparency adjustment parameters, and feeds the cutting plane parameters and transparency adjustment parameters back to the multi-layer adaptive ray projection rendering module to update the rendering results. The quantitative measurement and annotation module is used to perform distance measurement, angle calculation and volume annotation operations in the 3D visualization scene, overlay the measurement results on the rendering results, and back-transmit the visualization effect evaluation results obtained during the measurement process to the organ perception optimization transfer function adaptive optimization module to trigger the adaptive adjustment of the optimization transfer function.

2. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 1, characterized in that, The organ-sensing optimized transfer function adaptive optimization module includes: The tissue CT value statistical analysis unit is used to perform voxel-level statistical analysis on the input CT volume data and extract the CT value distribution characteristics of various tissue structures in the hepatobiliary system. The probability distribution modeling unit is used to probabilistically model the CT value distribution of four types of tissues, namely liver parenchyma, gallbladder, bile duct and blood vessel, based on a four-component Gaussian mixture model. The mean, variance and mixture weight parameters of each tissue category are obtained by iterative optimization through the expectation-maximization algorithm. An optimized transfer function parameter generation unit is used to generate a transparency mapping function and a color mapping function based on the probability distribution modeling results. The transparency mapping function determines the voxel transparency value based on the maximum posterior probability of each tissue category, and the color mapping function is based on a weighted mixture of the feature colors of each tissue according to the posterior probability.

3. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 2, characterized in that, The organ perception optimization transfer function adaptive optimization module also includes an adaptive adjustment unit, which receives the visualization effect evaluation results from the quantitative measurement annotation module. When the boundary clarity or contrast of the target tissue region is lower than the preset clarity threshold, it automatically adjusts the transparency adjustment factor and hybrid weight parameter of the corresponding tissue category to optimize the visualization effect.

4. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 1, characterized in that, The multi-layered adaptive ray casting rendering module includes: The ray generation unit is used to generate sampled rays that pass through the volume data space based on the current viewpoint position and viewing direction; An adaptive sampling unit is used to dynamically adjust the sampling step size based on the local gradient magnitude and the viewpoint distance. In regions where the gradient magnitude is greater than a preset gradient threshold, the sampling step size is reduced to capture detailed information, while in regions where the distance from the viewpoint exceeds a preset distance threshold, the sampling step size is increased to reduce computational overhead. The voxel interpolation unit is used to obtain the corresponding CT value from the volume data based on the sampling point location using the trilinear interpolation method; The color accumulation unit is used to accumulate the color and transparency contributions of each sampling point along the light direction using a forward accumulation method to generate the final pixel color value.

5. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 4, characterized in that, The multi-layer adaptive ray casting rendering module also includes a multi-layer detail control unit, which dynamically switches the rendering precision level according to the interaction speed and frame rate indicators, including high precision mode, medium precision mode and low precision mode. The high precision mode uses volume data with the original resolution and the minimum sampling step size, the medium precision mode uses volume data with 2x downsampling and 1.5x sampling step size, and the low precision mode uses volume data with 4x downsampling and 2x sampling step size.

6. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 1, characterized in that, The virtual anatomy interactive control module includes: The cutting plane control unit is used to respond to the user's cutting plane definition operation. It determines the cutting plane parameters by plane normal vector and a point on the plane. It supports the combined clipping of up to six cutting planes. The combined clipping uses a logical AND operation and only renders the area located on the effective side of all cutting planes. The transparency adjustment unit provides independent transparency control for four types of tissues: liver parenchyma, gallbladder, bile duct, and blood vessels. It uses a multiplicative correction method to multiply the user-set transparency correction coefficient by the original transparency value output by the optimized transfer function to obtain the final display transparency.

7. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 6, characterized in that, The virtual anatomy interactive control module also includes an interactive event processing unit, which is used to capture and parse the user's interactive operations, maintain the interactive state machine to track the current operating mode, and transmit the state information to other modules to coordinate system behavior.

8. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 1, characterized in that, The quantitative measurement and labeling module includes: The distance measurement unit is used to measure the Euclidean distance between any two points in three-dimensional space. It uses a sub-voxel-level positioning method to automatically lock onto the nearest tissue surface point through gradient search. Angle calculation unit is used to measure the angle formed by any three points in three-dimensional space, with the second point as the vertex of the angle, and to calculate the angle formed by the vector from the first point to the second point and the vector from the second point to the third point. The volume annotation unit provides two volume measurement methods: automatic volume calculation based on threshold segmentation and interactive volume calculation based on region growing.

9. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to claim 8, characterized in that, The quantitative measurement annotation module also includes a result visualization unit, which uses a layered rendering strategy to overlay the measurement results onto the volume rendering image. The result visualization unit uses depth testing and occlusion detection technology to automatically adjust the annotation position or use dashed lines to display the occluded part when the annotation is occluded by the organizational structure.

10. The interactive three-dimensional visualization reconstruction system for the hepatobiliary system based on volume rendering technology according to any one of claims 1 to 9, characterized in that, The system establishes a closed-loop feedback optimization mechanism. When the quantitative measurement and annotation module performs the measurement operation, it simultaneously evaluates the visualization quality index of the target tissue region. When the visualization quality index is lower than the preset quality threshold, the feedback signal is transmitted to the organ perception transfer function adaptive optimization module, which triggers the adaptive adjustment of the optimized transfer function parameters and updates the rendering results.

Images