CN110111386B - Method and system for automatically measuring structural angle in muscle tissue - Google Patents
Method and system for automatically measuring structural angle in muscle tissue Download PDFInfo
- Publication number
- CN110111386B CN110111386B CN201910389825.XA CN201910389825A CN110111386B CN 110111386 B CN110111386 B CN 110111386B CN 201910389825 A CN201910389825 A CN 201910389825A CN 110111386 B CN110111386 B CN 110111386B
- Authority
- CN
- China
- Prior art keywords
- muscle
- angle
- image
- fascia
- included angle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/45—For evaluating or diagnosing the musculoskeletal system or teeth
- A61B5/4519—Muscles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
Abstract
The invention discloses a method and a system for automatically measuring a structural angle in muscle tissue, wherein the method comprises the following steps: acquiring a muscle ultrasonic image in real time in muscle contraction movement, and preprocessing the muscle ultrasonic image to enhance image edge information; dividing fascia and muscle bundle regions in the muscle ultrasonic image by using the called normalized Radon transform algorithm, and acquiring a first included angle between a deep fascia and a preset horizontal axis; carrying out normalized Radon transformation on the fascicle area, extracting a plurality of peak points in a Radon transformation matrix, and carrying out weighted average on the positions and the directions of all the muscle fibers to obtain a second included angle between the muscle fibers and a horizontal axis; and summing the first included angle and the second included angle to obtain a structural angle in the muscle tissue, wherein the structural angle is a muscle pinnate angle. The invention can naturally, smoothly, robustly and automatically measure the structural angle in the muscle tissue, so that the measurement result is closer to the real motion state, and the structural function of the muscle can be reflected better.
Description
Technical Field
The invention relates to the technical field of measurement of muscle structural angles, in particular to an automatic measurement method and system of structural angles in muscle tissues.
Background
Muscles are important tissues constituting the human body, and their main functions are to generate contraction and control the movement of the body. The muscle structure has a major influence on the way muscle force is transmitted to the tendons and bones. The structural angle of muscle, such as the pinnate angle of muscle, is an important index for skeletal muscle research. The muscle structural angle is highly related to the muscle contraction strength, so that doctors can know the biomechanical characteristics of muscles, better evaluate the muscle functions and pathological states, and provide effective basis for clinical diagnosis and rehabilitation evaluation.
Most of the existing measurement methods for the structural angle in the muscle tissue adopt manual drawing, the manual measurement method is tedious, time-consuming, poor in repeatability and quite subjective, and the analysis of the muscle function in a dynamic task is difficult to realize. The general automatic measurement method also has certain defects, which are mainly embodied as the following three points: the first point is that the structural angle dynamic curve measured by the algorithm has low smoothness. The original signals with dynamically changed angles detected by the methods are not smooth, have fluctuation of high-frequency jitter, do not conform to the rule of contraction motion of a human muscle system, and influence the measurement precision. The second point is that the algorithm is less robust. These algorithms are often only suitable for ultrasound images of a certain type of muscle, and have serious degradation and poor generalization capability on individual data sets. The third point is that the algorithm is less automated. The whole operation flow is complicated, an initial point needs to be manually set or parameters need to be manually adjusted, and the automation degree is not high.
Therefore, the prior art has yet to be improved.
Disclosure of Invention
The technical problem to be solved by the present invention is to provide a method and a system for automatically measuring a structural angle in a muscle tissue, aiming at solving the problems of low precision, low robustness, low automation degree, etc. of the method for measuring a structural angle of a muscle in the prior art.
The technical scheme adopted by the invention for solving the technical problem is as follows:
a method of automated measurement of structural angulation in muscle tissue, wherein the method comprises:
acquiring a muscle ultrasonic image in real time in muscle contraction movement, and preprocessing the muscle ultrasonic image to enhance image edge information of the muscle ultrasonic image;
calling a preset normalized Radon transform algorithm to divide the fascia and muscle bundle areas in the muscle ultrasonic image, and acquiring a first included angle between the deep fascia and a preset horizontal axis in the muscle ultrasonic image;
performing normalized Radon transform on the muscle bundle region, extracting a plurality of peak points in a Radon transform matrix for forming a plurality of bundles of muscle fibers in Euclidean space, and performing weighted average on the positions and directions of all the muscle fibers to obtain a second included angle between the muscle fibers and a preset horizontal axis in the muscle ultrasonic image;
and summing the first included angle and the second included angle to obtain a structural angle in the muscle tissue, wherein the structural angle is a muscle pinnate angle.
The method for automatically measuring the structural angle in the muscle tissue comprises the following steps of obtaining an ultrasonic image by adopting an ultrasonic scanner, wherein the long axis of a probe of the ultrasonic scanner is vertical to the long axis direction of the muscle.
The method for automatically measuring structural angles in muscle tissues, wherein the preprocessing the muscle ultrasound image to enhance image edge information of the muscle ultrasound image comprises:
and acquiring each salient region in the muscle ultrasonic image, calculating a histogram of each salient region, redistributing the image brightness value of each salient region, and enhancing the local detail and edge information of the image.
The automatic measurement method of the structural angle in the muscle tissue, wherein the invoking of the preset normalized radon transform algorithm to divide the fascia and the fascial tract regions in the muscle ultrasound image comprises:
carrying out normalized Radon transformation on the muscle ultrasonic image at a preset angle interval;
extracting a peak point on the Redon transformation matrix, and taking the position of the peak point as a center and the line width and the angle fluctuation range of the fascia as the length and the width to obtain a target matrix area;
in the Redon transformation matrix, clearing an area formed by taking the peak point position as a center, taking twice of the distance between the fasciae as a length and the width of the matrix as a width;
acquiring a plurality of peak points in the target matrix area, and identifying fascia by using the position and direction information of the peak points;
repeating the above process to detect multiple fascia, and completing division of fascia and muscle bundle regions.
The automatic measurement method of the structural angle in the muscle tissue, wherein the formula of the normalized Redon transform is as follows:
R(ρ，θ)＝∫∫ D (I(x，y)-mean(I))δ(ρ-xcosθ-ysinθ)dxdy；
wherein NR (ρ, θ) = mapminmax (R (ρ, θ)); r (ρ, θ) represents the image after standard radon transform, mean (I) represents the mean of the gray levels of the image, and mapminmax (R (ρ, θ)) represents normalizing the matrix to between 0 and 1.
The method for automatically measuring the structural angle in the muscle tissue, wherein the acquiring a first included angle between the deep fascia and a preset horizontal axis in the muscle ultrasound image comprises:
selecting a preset number of peak points in the target area matrix, and performing weighted average on the position and the direction of the fascia by taking the brightness values of the selected peak points as weights;
recording an average included angle between the deep fascia and a horizontal axis in the muscle ultrasound image, wherein the average included angle is the first included angle.
The automatic measurement method of the structural angle in the muscle tissue, wherein the calculation formula of the weighted average is as follows:
wherein, the first and the second end of the pipe are connected with each other,
represents the average direction of all muscle fibers, θ i Representing the direction of each bundle of muscle fibres, omega i Representing the weight of each bundle of muscle fibers.
An automatic measurement system of a structural angle in muscle tissue, wherein the system comprises:
the image acquisition and enhancement module is used for acquiring a muscle ultrasonic image in real time in the muscle contraction movement and preprocessing the muscle ultrasonic image so as to enhance the image edge information of the muscle ultrasonic image;
the first included angle acquisition module is used for calling a preset normalized Radon transform algorithm to divide the fascia and muscle bundle regions in the muscle ultrasonic image and acquiring a first included angle between the deep fascia and a preset horizontal axis in the muscle ultrasonic image;
a second included angle obtaining module, configured to perform normalized radon transform on the muscle bundle region, extract a plurality of peak points in a radon transform matrix, the peak points being used to form a plurality of bundles of muscle fibers in an euclidean space, and perform weighted average on positions and directions of all the muscle fibers, to obtain a second included angle between the muscle fibers and a horizontal axis preset in the muscle ultrasound image;
and the angle summation calculation module is used for carrying out summation calculation on the first included angle and the second included angle to obtain a structural angle in the muscle tissue, wherein the structural angle is a muscle pinnate angle.
The automatic measurement system for the structural angle in the muscle tissue is characterized in that the muscle ultrasonic image is an image obtained by adopting an ultrasonic scanner, and the long axis of a probe of the ultrasonic scanner is vertical to the long axis direction of the muscle.
The automatic measurement system for the structural angle in the muscle tissue, wherein the first included angle acquisition module comprises a partitioning and positioning sub-module, and the partitioning and positioning sub-module is used for: carrying out normalized Radon transformation on the muscle ultrasonic image at a preset angle interval;
extracting a peak point on the Redon transformation matrix, and taking the position of the peak point as a center and the line width and the angle fluctuation range of the fascia as the length and the width to obtain a target matrix area;
in the Redon transformation matrix, clearing an area which is formed by taking the peak point position as the center, taking twice of the distance between the fasciae as the length and taking the width of the matrix as the width;
and acquiring a plurality of peak points in the target matrix area, identifying the fascia by using the position and direction information of the peak points, detecting a plurality of fascias, and completing division of the fascia and the fascial area.
The invention has the beneficial effects that: according to the invention, the muscle ultrasonic image in the muscle contraction process is obtained in real time, and the normalized Radon transformation is carried out on the preprocessed image, so that the image edge effect can be effectively eliminated, the problem of highlight of an image diagonal is solved, the measurement robustness of the structural angle in the muscle image is stronger, and the accuracy is higher.
Drawings
FIG. 1 is a flow chart of a preferred embodiment of the method for automated measurement of structural angulation in muscle tissue provided by the present invention.
Fig. 2 is a schematic diagram of an ultrasound image of muscle in the automatic measurement method for structural angles in muscle tissue provided by the present invention.
Fig. 3 is a detailed flow chart of the method for automatically measuring structural angles in muscle tissue provided by the present invention.
FIG. 4 is a comparison graph of the ultrasound image of a muscle with the enhancement before and after the method for automatically measuring the structural angle in the muscle tissue provided by the present invention.
Fig. 5 is an enhanced image of a muscle ultrasound image and its corresponding raldon transform matrix in the method for automatic measurement of structural angles in muscle tissue according to the present invention.
FIG. 6 is a schematic diagram of peak points extracted during the process of dividing the fascia and fascial regions in the automatic measurement method for structural angles in muscle tissues provided by the present invention.
FIG. 7 is a schematic diagram of muscle fibers extracted by normalized Radon transform in the automated method for measuring structural angles in muscle tissue according to the present invention.
Fig. 8 is a functional schematic diagram of an automatic measurement system for structural angles in muscle tissue according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Muscles are important tissues constituting the human body, and their main functions are to produce contraction and control the movement of the body. It has been found that the muscle structure has a major influence on the way muscle force is transmitted to the tendons and bones [1] . The structural angle of muscle, such as the pinnate angle of muscle, is an important index for skeletal muscle research. The muscle structural angle is highly related to the muscle contraction strength, so that doctors can know the biomechanical characteristics of muscles, better evaluate the muscle functions and pathological states, provide effective basis for clinical diagnosis and rehabilitation evaluation, and serve rehabilitation medicine or physical science. However, the constitution of the muscle is very complicated, and how to accurately quantify the structure of the muscle during dynamic contraction is a hotspot and difficulty in the research of motor function rehabilitation and motor medicine.
The muscle ultrasonic imaging is a non-invasive, non-radiative, low-cost, real-time technique for displaying normal tissues and pathological muscle tissues. Two-dimensional ultrasound images have been widely used to study changes in the morphological structure of muscle movements, providing an effective means for clinical diagnosis and rehabilitation assessment, due to the ability to measure in any posture and to provide real-time dynamic images of structural target changes as the muscles contract. The muscle function and pathological state are evaluated by acquiring muscle images by using an ultrasonic instrument and analyzing the structural parameters of the muscle. The structural angle in the muscle tissue is used as an important and intuitive parameter for evaluating the functional state of the muscle, and how to measure the structural angle of the muscle naturally, smoothly, robustly and accurately is important.
At present, there are three methods for automatically measuring the structural angle in the muscle ultrasound image by using an image processing technology: the first is feature tracking based on multiple ultrasonic images, the second is feature detection based on a single ultrasonic image, and the third is end-to-end detection based on deep learning.
The feature tracking method is to manually delineate a region of interest on a first frame image, and then dynamically track a selected object between adjacent frames by using a tracking algorithm, such as a Lucas-Kanade optical flow algorithm (a Lucas-Kanade optical flow estimation algorithm is a two-frame differential optical flow estimation algorithm) or a cross-correlation algorithm. Tracking features in ultrasound images is challenging because it must deal with problems of muscle deformation, speckle noise, and significant changes in the appearance or intensity of the tracked object in the region of interest. Currently, although some tracking algorithms perform well on some specific tasks, they still exhibit cumulative tracking errors due to the problems of non-uniform deformation of the region of interest between successive images, noise interference and even disappearance; moreover, the tracking algorithm needs to manually select the target tracking point of the first frame, and the semi-automatic interaction mode can be deviated due to the difference of operator experience.
The method of feature detection is to identify structural objects, such as fascia and muscle fibers, by locating these linear structures in the muscle ultrasound image, assuming they are linear. On a muscle ultrasound image, the fascia is continuously highly echogenic, while the muscle fibers are usually unevenly distributed in a linear structure. By detecting muscle fibers using hough transform or radon transform, measurement of the structural angle of the muscle can be achieved. Although a general feature detection method can determine a representative beam direction through hough transform or radon transform, speckle noise and the appearance of intramuscular blood vessels often mask important features of structural targets on an ultrasonic image, and the difficulty of accurately measuring structural angles is increased.
The detection method of deep learning refers to an architecture neural network model, and the direction of a local structural target is predicted by learning structural target information in a well-defined region. Although the deep learning method is applied to the performance of the skeletal muscle fiber direction in the whole region, which is shown in an acceptable error range, due to the data-driven characteristic, the model needs to train a large amount of data, and the requirement on computing power resources and training time is high. Moreover, the technology is not mature at present, and the dynamic measurement of the muscle structural angle still has larger deviation.
In order to solve the above problem, the present embodiment provides a method for automatically measuring a structural angle in muscle tissue, as shown in fig. 1 in particular, the method includes the following steps:
s100, acquiring a muscle ultrasonic image in real time in muscle contraction movement, and preprocessing the muscle ultrasonic image to enhance image edge information of the muscle ultrasonic image;
step S200, calling a preset normalized Radon transform algorithm to divide fascia and muscle bundle areas in the muscle ultrasonic image, and acquiring a first included angle between a deep fascia and a preset horizontal axis in the muscle ultrasonic image;
step S300, carrying out normalized Radon transformation on the muscle bundle region, extracting a plurality of peak points which are used for forming a plurality of bundles of muscle fibers in Euclidean space in a Radon transformation matrix, carrying out weighted average on the positions and the directions of all the muscle fibers, and obtaining a second included angle between the muscle fibers and a preset horizontal axis in the muscle ultrasonic image;
and S400, summing the first included angle and the second included angle to obtain a structural angle in the muscle tissue, wherein the structural angle is a muscle pinnate angle.
In the embodiment, the muscle ultrasonic image in the muscle contraction process is obtained through implementation, the normalized Radon transform is carried out on the preprocessed image, the image edge effect can be effectively eliminated, the problem of highlight of an image diagonal line is solved, the measurement robustness of the structural angle in the muscle image is enabled to be stronger, the accuracy is higher, and the structural angle in the muscle tissue can be naturally, smoothly, robustly and accurately measured with high automation by adopting a multi-muscle fiber weighted average algorithm, so that the measurement result is closer to the real motion state, and the structural function of the whole muscle can be reflected better.
Specifically, in this embodiment, first, when the muscle contracts, the ultrasound image generated by the muscle contraction is recorded, and of course, the signals recorded in this process include not only the ultrasound image but also dynamic changes of output torque values or signals of myoelectricity, joint angles, and the like. There are many instruments and methods for recording ultrasound images and outputting signals, and in this embodiment, a real-time B-mode ultrasound scanner with a 10MHz electronic linear array probe is preferred to acquire muscle ultrasound images. In particular implementations, the long axis of the ultrasound probe should be perpendicular to the long axis of the muscle, 40% distal to the knee joint (measured from the anterior superior iliac spine to the superior patellar margin). The ultrasound probe is secured by a custom designed foam container with securing straps and a large amount of ultrasound gel is used during muscle contraction to ensure acoustic coupling between the probe and the skin. The ultrasonic probe can be adjusted to optimize the contrast of the muscle bundles in the ultrasonic image, and the position is marked to ensure that the probe is placed at the same position every time, so that the accuracy of image acquisition is improved.
Referring to fig. 2, fig. 2 is a muscle ultrasound image, and fig. 2 shows that the muscle layer of the skeletal muscle includes a deep fascia, a superficial fascia and a muscle bundle region, a muscle bundle line L1 and a deep fascia line L2 can be drawn according to the arrangement direction of muscle fibers and the direction of the deep fascia, and an included angle θ between the two lines is defined as a pinnate angle, which is a structural angle required to be measured by the present invention. Specifically, as shown in fig. 3, the measurement method in this embodiment first preprocesses the muscle ultrasound image to enhance the image edge information, thereby facilitating the segmentation of the deep, superficial fascia and muscle bundle regions. Next, the present embodiment uses the normalized radon transform to precisely locate and divide the deep and shallow fascia, and calculates an angle θ 1 (a first angle) between the deep fascia and a horizontal line (i.e., an X-axis on the muscle ultrasound image). Then, normalized radon transform is performed again on the divided muscle bundle region, a large number of peak points in a radon transform matrix are extracted, a plurality of bundles of muscle fibers are formed in euclidean space, the directions and positions of all the muscle fibers are weighted and averaged, an included angle θ 2 (a second included angle) between the included angle θ 2 and a horizontal line ((namely, an X axis on the muscle ultrasound image)) is recorded, and the pinnate angle θ (θ = θ 1+ θ 2) of the muscle ultrasound image can be measured by summing the included angle θ 1 and the included angle θ 2 according to a formula.
Further, since the muscle ultrasound image contains fascia, fascial texture features, and noise information. Important textural features are susceptible to high speckle noise and interference from intramuscular blood vessels, resulting in increased difficulty in accurately locating each tissue. Therefore, the present embodiment needs enhancement processing of the muscle ultrasound image. Preferably, the present embodiment employs a Contrast-Limited Adaptive Histogram Equalization (CLAHE), which is an image processing technique for improving the Contrast of an image. Specifically, each salient region in the muscle ultrasonic image is obtained, a histogram of each salient region is calculated, the image brightness value of each salient region is redistributed, and the local detail and edge information of the image are enhanced. Meanwhile, each pixel neighborhood on the image is subjected to contrast limitation to overcome noise interference, as shown in fig. 4, a in fig. 4 is an original muscle ultrasound image, and b is a muscle ultrasound image enhanced by CLAHE. In the embodiment, the CLAHE can effectively inhibit noise, reduce errors caused by the noise, enhance the texture characteristics of the muscle bundles, and facilitate the improvement of the positioning and identifying capability of the Radon transform on the structural target.
In particular, on a muscle ultrasound image, the fascia usually appears as a continuous high-echo band with an orientation close to zero degrees, while the muscle fibers usually appear as unevenly distributed threadlike structures with an orientation that is also traversable. Thus, the present embodiment employs the radon transform for detection of the continuous fascia and isotropic muscle fibers. The formula of the standard radon transform in the euclidean space is:
R(ρ，θ)＝∫∫ D I(x，y)δ(ρ-xcosθ-ysinθ)dx dy
wherein, I (x, y) represents the image gray scale at the (x, y) position, delta is a Dirac delta function, rho is the distance from a straight line to the center of the image, theta is the included angle between the x axis and the line, and D is the image grid. The standard radon transform represents the gray scale integral of the whole image plane in the projection direction, and is widely applied to the extraction of straight lines and the identification of boundaries because peak points on a radon transform matrix can represent the characteristics of the straight lines. However, since the integration path lengths are different in different projection directions, edge effects are easily generated, and as shown in fig. 5, it is not possible to determine which direction represented by the peak point is the fascia direction. To eliminate the edge effect, the present embodiment proposes a new normalized raldon transform:
R(ρ，θ)＝∫∫ D (I(x，y)-mean(I))δ(ρ-xcosθ-ysinθ)dxdy
NR(ρ，θ)＝mapminmax(R(ρ，θ))；
where R (ρ, θ) represents the image after standard radon transform, mean (I) represents the mean of the gray levels of the image, and mapminmax (R (ρ, θ)) represents the normalization of the matrix to between 0 and 1. In this embodiment, the problem of the general highlight of the diagonal lines in the radon transform matrix is solved by subtracting the image grayscale mean from the original image, and then the image is normalized radon transformed at a preset angle interval (for example, 0.1 degree) according to the prior knowledge of the fascia direction range. In order to accurately divide the fascia and muscle bundle regions, the present embodiment further includes the following steps:
extracting a peak point on the Radon transformation matrix, and taking the position of the peak point as a center and the line width and the angle fluctuation range of the fascia as the length and the width to obtain a target matrix area;
in the Redon transformation matrix, clearing an area which is formed by taking the peak point position as the center, taking twice of the distance between the fasciae as the length and taking the width of the matrix as the width;
acquiring a plurality of peak points in the target matrix area, and identifying fascia by using the position and direction information of the peak points;
repeating the above process to detect multiple fascia and complete the division of fascia and fascial regions.
In this embodiment, the fascia and the muscle bundle region in the muscle ultrasound image can be accurately divided through the above steps, as shown in fig. 6, a diagram in fig. 6 is an extracted fascia peak point (a black point in fig. 6), and b diagram is a straight line corresponding to the peak point, so that accurate positioning of the fascia and the muscle bundle region is realized.
Further, the luminance value on the matrix of the Redon transform represents the accumulated result of each straight line on the original image mapped onto the Redon space. The larger the value, the greater the probability that the inverse transform back into euclidean space is a straight line. Therefore, the present invention uses this information to select a predetermined number of peak points (e.g., 50 peak points) in the target matrix region, weight-average the position and direction with their respective luminance values as weights, and record the average angle θ 1 between the deep fascia and the horizontal line (X-axis in the image). The weighted average is calculated as follows:
wherein the content of the first and second substances,
represents the average direction of all muscle fibers, θ i Representing the direction of each bundle of muscle fibres, ω i Representing the weight of each bundle of muscle fibers. The embodiment can accurately identify the dominant direction of the fascia, and realizes the automatic division of deep and shallow fascia and muscle bundle regions.
Further, in this embodiment, according to a priori knowledge of the range of the muscle fiber direction, normalized radon transform may be performed on the divided muscle bundle region at certain angular intervals (for example, 0.1 degree), a plurality of peak points (for example, 1000 peaks) in the radon transform matrix are extracted, and a plurality of bundles of muscle fibers are formed on euclidean space, as shown in fig. 7, where a in fig. 7 is the extracted muscle fiber peak point (black point in fig. 7), and b is a straight line corresponding to the peak point. The position and orientation of all muscle fibers are then weighted averaged and recorded at an angle θ 2 to the horizontal (X-axis in the image). Then adding theta 1 and theta 2, the muscle pinnate angle of each frame of muscle ultrasonic image, namely the structural angle in muscle tissue, can be accurately calculated.
In the embodiment, the ultrasonic image is processed by using a multi-muscle fiber weighted average algorithm to dynamically study the structural angle of the muscle and the change of related morphological parameters (such as muscle thickness, muscle fiber angle, muscle bundle length, physiological cross-sectional area of the muscle, and the like), so that the measurement precision of the structural angle is effectively improved, and the robustness is high. The automatic measurement method for the structural angle in the muscle tissue provided by the embodiment can overcome high speckle noise and the interference of intramuscular blood vessels, effectively divide fascia and muscle bundle areas, and realize full-automatic accurate measurement of the dynamic change of the structural angle in the muscle contraction process. Compared with the existing measuring method, the measuring result of the embodiment is more consistent with the result manually drawn by an expert, the correlation with the muscle strength is higher, and the dynamic curve of the structural angle is natural and smooth; the algorithm can be applied to a plurality of groups of data sets of different types, the robustness is high, the accurate measurement of the structural angle can be completed without manually adjusting parameters, and the automation degree is high.
Based on the above embodiment, the present invention also provides an automatic measurement system of a structural angle in muscle tissue, as shown in fig. 8, the system comprising: an image acquisition and enhancement module 810, a first angle acquisition module 820, a second angle acquisition module 830, and an angle summation calculation module 840.
Specifically, the image obtaining and enhancing module 810 is configured to obtain a muscle ultrasound image in real time during a muscle contraction motion, and preprocess the muscle ultrasound image to enhance image edge information of the muscle ultrasound image;
a first included angle obtaining module 820, configured to invoke a preset normalized raldon transform algorithm to divide the fascia and the muscle bundle region in the muscle ultrasound image, and obtain a first included angle between the deep fascia and a preset horizontal axis in the muscle ultrasound image;
a second included angle obtaining module 830, configured to perform normalized radon transform on the muscle bundle region, extract a plurality of peak points in a radon transform matrix, where the peak points are used to form a plurality of bundles of muscle fibers in euclidean space, and perform weighted average on the positions and directions of all the muscle fibers to obtain a second included angle between the muscle fibers and a horizontal axis preset in the muscle ultrasound image;
the angle summation calculation module 840 is configured to perform summation calculation on the first included angle and the second included angle to obtain a structural angle in a muscle tissue, where the structural angle is a muscle pinnate angle.
Preferably, the embodiment prefers a real-time B-mode ultrasonic scanner of a 10MHz electronic linear array probe to acquire the muscle ultrasonic image. In particular implementations, the long axis of the ultrasound probe should be perpendicular to the long axis of the muscle, 40% distal to the knee joint (measured from the anterior superior iliac spine to the superior patellar margin).
Preferably, in this embodiment, the first included angle obtaining module 820 includes a dividing and positioning sub-module, and the dividing and positioning sub-module is configured to: carrying out normalized Radon transformation on the muscle ultrasonic image at a preset angle interval; extracting a peak point on the Redon transformation matrix, and taking the position of the peak point as a center and the line width and the angle fluctuation range of the fascia as the length and the width to obtain a target matrix area; in the Redon transformation matrix, clearing an area formed by taking the peak point position as a center, taking twice of the distance between the fasciae as a length and the width of the matrix as a width; and acquiring a plurality of peak points in the target matrix region, identifying fascia by using the position and direction information of the peak points, detecting a plurality of fascias, and completing division of the fascia and the fasciae region.
In the embodiment, the muscle ultrasonic image in the muscle contraction process is obtained in real time, the normalized Radon transform is performed on the preprocessed image, the image edge effect can be effectively eliminated, the problem of highlight of an image diagonal line is solved, the measurement robustness of the structural angle in the muscle image is higher, the accuracy is higher, and the structural angle in the muscle tissue can be naturally, smoothly, robustly and accurately measured with high automation by adopting a multi-muscle fiber weighted average algorithm, so that the measurement result is closer to the real motion state, and the structural function of the whole muscle can be reflected better.
In summary, the present invention discloses a method and a system for automatically measuring structural angles in muscle tissue, wherein the method comprises: acquiring a muscle ultrasonic image in real time in muscle contraction movement, and preprocessing the muscle ultrasonic image to enhance image edge information of the muscle ultrasonic image; calling a preset normalized Radon transform algorithm to divide the fascia and muscle bundle areas in the muscle ultrasonic image, and acquiring a first included angle between the deep fascia and a preset horizontal axis in the muscle ultrasonic image; performing normalized Radon transform on the muscle bundle region, extracting a plurality of peak points in a Radon transform matrix for forming a plurality of bundles of muscle fibers in Euclidean space, and performing weighted average on the positions and directions of all the muscle fibers to obtain a second included angle between the muscle fibers and a preset horizontal axis in the muscle ultrasonic image; and summing the first included angle and the second included angle to obtain a structural angle in the muscle tissue, wherein the structural angle is a muscle pinnate angle. The invention can naturally, smoothly, robustly and automatically measure the structural angle in the muscle tissue, so that the measurement result is closer to the real motion state, and the structural function of the muscle can be reflected better.
It will be understood that the invention is not limited to the examples described above, but that modifications and variations will occur to those skilled in the art in light of the above teachings, and that all such modifications and variations are considered to be within the scope of the invention as defined by the appended claims.
Claims (7)
1. A method of automatically measuring a structural angle in muscle tissue, the method comprising:
acquiring a muscle ultrasonic image in real time in muscle contraction movement, and preprocessing the muscle ultrasonic image to enhance image edge information of the muscle ultrasonic image;
calling a preset normalized Radon transform algorithm to divide the fascia and muscle bundle areas in the muscle ultrasonic image, and acquiring a first included angle between the deep fascia and a preset horizontal axis in the muscle ultrasonic image;
performing normalized Radon transform on the muscle bundle region, extracting a plurality of peak points in a Radon transform matrix for forming a plurality of bundles of muscle fibers in Euclidean space, performing weighted average on the positions and directions of all the muscle fibers, and acquiring a second included angle between the muscle fibers and a preset horizontal axis in the muscle ultrasonic image;
summing the first included angle and the second included angle to obtain a structural angle in the muscle tissue, wherein the structural angle is a muscle pinnate angle;
the step of calling a preset normalized Radon transform algorithm to divide the fascia and muscle bundle regions in the muscle ultrasound image comprises the following steps:
carrying out normalized Radon transformation on the muscle ultrasonic image at a preset angle interval;
extracting a peak point on the Redon transformation matrix, and taking the position of the peak point as a center and the line width and the angle fluctuation range of the fascia as the length and the width to obtain a target matrix area;
in the Redon transformation matrix, clearing an area formed by taking the peak point position as a center, taking twice of the distance between the fasciae as a length and the width of the matrix as a width;
acquiring a plurality of peak points in the target matrix area, and identifying fascia by using the position and direction information of the peak points;
repeating the process, detecting a plurality of fascia, and completing division of the fascia and the muscle bundle regions; the formula of the normalized Redon transform is as follows:
wherein the content of the first and second substances,
，
representing the image after a standard radon transform,
represents the mean value of the gray levels of the image,
represents normalizing the matrix to between 0 and 1;
the acquiring a first angle between the deep fascia and a preset horizontal axis in the muscle ultrasound image comprises:
selecting a preset number of peak points in the target matrix area, and performing weighted average on the position and the direction of the fascia by taking the brightness values of the selected peak points as weights;
recording an average included angle between the deep fascia and a horizontal axis in the muscle ultrasound image, wherein the average included angle is the first included angle.
2. The method of claim 1, wherein the muscle ultrasound image is an image obtained using an ultrasound scanner having a probe with a long axis perpendicular to a long axis direction of the muscle.
3. The method of claim 1, wherein the pre-processing the muscle ultrasound image to enhance image edge information of the muscle ultrasound image comprises:
and acquiring each salient region in the muscle ultrasonic image, calculating a histogram of each salient region, redistributing the image brightness value of each salient region, and enhancing the local detail and edge information of the image.
4. The method of claim 1, wherein the weighted average is calculated by the formula:
wherein, the first and the second end of the pipe are connected with each other,
representing the average direction of all the muscle fibers,
representing the direction of each bundle of muscle fibers,
representing the weight of each bundle of muscle fibers.
5. An automatic measurement system of structural angles in muscle tissue based on the automatic measurement method of structural angles in muscle tissue according to any one of claims 1 to 4, characterized in that the system comprises:
the image acquisition and enhancement module is used for acquiring a muscle ultrasonic image in real time in muscle contraction motion and preprocessing the muscle ultrasonic image so as to enhance the image edge information of the muscle ultrasonic image;
the first included angle acquisition module is used for calling a preset normalized Radon transform algorithm to divide the fascia and muscle bundle areas in the muscle ultrasonic image and acquiring a first included angle between the deep fascia and a preset horizontal axis in the muscle ultrasonic image;
a second included angle obtaining module, configured to perform normalized radon transform on the muscle bundle region, extract a plurality of peak points in a radon transform matrix, where the peak points are used to form a plurality of bundles of muscle fibers in a euclidean space, and perform weighted average on positions and directions of all the muscle fibers to obtain a second included angle between the muscle fibers and a horizontal axis preset in the muscle ultrasound image;
and the angle summation calculation module is used for carrying out summation calculation on the first included angle and the second included angle to obtain a structural angle in the muscle tissue, wherein the structural angle is a muscle pinnate angle.
6. The system of claim 5, wherein the ultrasound image of the muscle is an image obtained using an ultrasound scanner having a probe with a long axis perpendicular to the long axis of the muscle.
7. The system of claim 6, wherein the first angle acquisition module comprises a divide and locate sub-module configured to: carrying out normalized Radon transformation on the muscle ultrasonic image at a preset angle interval;
extracting a peak point on the Redon transformation matrix, and taking the position of the peak point as a center and the line width and the angle fluctuation range of the fascia as the length and the width to obtain a target matrix area;
in the Redon transformation matrix, clearing an area which is formed by taking the peak point position as the center, taking twice of the distance between the fasciae as the length and taking the width of the matrix as the width;
and acquiring a plurality of peak points in the target matrix area, identifying the fascia by using the position and direction information of the peak points, detecting a plurality of fascias, and completing division of the fascia and the fascial area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910389825.XA CN110111386B (en) | 2019-05-10 | 2019-05-10 | Method and system for automatically measuring structural angle in muscle tissue |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910389825.XA CN110111386B (en) | 2019-05-10 | 2019-05-10 | Method and system for automatically measuring structural angle in muscle tissue |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110111386A CN110111386A (en) | 2019-08-09 |
| CN110111386B true CN110111386B (en) | 2023-01-24 |
Family
ID=67489427
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910389825.XA Active CN110111386B (en) | 2019-05-10 | 2019-05-10 | Method and system for automatically measuring structural angle in muscle tissue |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110111386B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110930394B (en) * | 2019-11-29 | 2021-07-16 | 深圳先进技术研究院 | Method and terminal equipment for measuring slope and pinnate angle of muscle fiber bundle line |
| CN111353979B (en) * | 2020-02-27 | 2023-10-17 | 成都恒创新星科技有限公司 | Mobile computer vision facula detection positioning method |
| CN112085711B (en) * | 2020-08-25 | 2023-12-22 | 山东科技大学 | Method for automatically tracking muscle feather angle by combining convolutional neural network and Kalman filtering |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004078698A (en) * | 2002-08-20 | 2004-03-11 | Japan Science & Technology Corp | Simulation system using human body model including muscle |
| CN102920436A (en) * | 2012-10-25 | 2013-02-13 | 深圳先进技术研究院 | Processing method of muscle image by utilizing Hough transform |
| CN108742705A (en) * | 2018-04-10 | 2018-11-06 | 深圳大学 | A kind of supersonic imaging apparatus and method of real-time detection muscle morphological parameters |
| CN108961278A (en) * | 2018-06-20 | 2018-12-07 | 深圳市旭东数字医学影像技术有限公司 | The method and its system of abdominal wall muscle segmentation based on image data |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009107500A1 (en) * | 2008-02-25 | 2009-09-03 | 株式会社 日立メディコ | Medical image processing device, medical image processing method, and program |
-
2019
- 2019-05-10 CN CN201910389825.XA patent/CN110111386B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004078698A (en) * | 2002-08-20 | 2004-03-11 | Japan Science & Technology Corp | Simulation system using human body model including muscle |
| CN102920436A (en) * | 2012-10-25 | 2013-02-13 | 深圳先进技术研究院 | Processing method of muscle image by utilizing Hough transform |
| CN108742705A (en) * | 2018-04-10 | 2018-11-06 | 深圳大学 | A kind of supersonic imaging apparatus and method of real-time detection muscle morphological parameters |
| CN108961278A (en) * | 2018-06-20 | 2018-12-07 | 深圳市旭东数字医学影像技术有限公司 | The method and its system of abdominal wall muscle segmentation based on image data |
Non-Patent Citations (2)
| Title |
|---|
| "归一化 Radon 变换和统计测度的超声图像肌束方向自动跟踪";李乔亮等;《中国图像图形学报》;20130930;第18卷(第9期);第1-3节 * |
| "腓肠肌羽状角的超声自动测量";潘庆亚等;《南方医科大学学报》;20151231;第35卷(第8期);第1-3节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110111386A (en) | 2019-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11633169B2 (en) | Apparatus for AI-based automatic ultrasound diagnosis of liver steatosis and remote medical diagnosis method using the same | |
| CN110111386B (en) | Method and system for automatically measuring structural angle in muscle tissue | |
| Loizou et al. | Snakes based segmentation of the common carotid artery intima media | |
| US8787622B2 (en) | Biometric authentication device, biometric authentication method, and recording medium | |
| Suhling et al. | Myocardial motion analysis from B-mode echocardiograms | |
| US9087397B2 (en) | Method and apparatus for generating an image of an organ | |
| JPH08206117A (en) | Ultrasonic diagnostic apparatus | |
| US20080077011A1 (en) | Ultrasonic apparatus | |
| US20080095417A1 (en) | Method for registering images of a sequence of images, particularly ultrasound diagnostic images | |
| US20230359057A1 (en) | Motion-compensated laser speckle contrast imaging | |
| US11455720B2 (en) | Apparatus for ultrasound diagnosis of liver steatosis using feature points of ultrasound image and remote medical-diagnosis method using the same | |
| JP2009512527A (en) | Image registration method, algorithm for performing the image registration method, program for registering an image using the algorithm, and biomedical image handling method for reducing image artifacts due to object movement | |
| EP0533782B1 (en) | Automated method for digital image quantitation | |
| US9014452B2 (en) | Orientation-aware average intensity histogram to indicate object boundary depth in ultrasound images | |
| EP2637570A2 (en) | System and method of ultrasound image processing | |
| CN113066064B (en) | Cone beam CT image biological structure identification and three-dimensional reconstruction system based on artificial intelligence | |
| CN111462156B (en) | Image processing method for acquiring corneal vertex | |
| CN104732520A (en) | Cardio-thoracic ratio measuring algorithm and system for chest digital image | |
| JP3696616B2 (en) | Ultrasonic diagnostic equipment | |
| CN106456084A (en) | Ultrasound imaging apparatus | |
| CN110276772B (en) | Automatic positioning method and system for structural elements in muscle tissue | |
| JP2005137936A (en) | Ultrasonic diagnostic equipment | |
| CN110930394B (en) | Method and terminal equipment for measuring slope and pinnate angle of muscle fiber bundle line | |
| EP2693397A1 (en) | Method and apparatus for noise reduction in an imaging system | |
| CN112330603A (en) | System and method for estimating motion of target in tissue based on soft tissue surface deformation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |