Vital signs monitoring method for critically ill patients based on image processing
By analyzing the pupil and iris characteristics of critically ill patients and calculating the static and dynamic pupillary sign coefficients, the problem of non-objective pupillary quantification in existing technologies is solved, and the efficiency and accuracy of vital sign monitoring are improved.
Patent Information
- Application Number
- CN202411204313.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-30
AI Technical Summary
Current technologies lack objectivity in quantitative analysis of pupil morphology in critically ill patients, have low efficiency in monitoring vital signs, and make it difficult to quickly and accurately assess patients' conditions.
By acquiring images and frame sequences of the patient's eyeballs, the relative position, shape, size, and offset parameters of the pupil and iris are analyzed to calculate static and dynamic pupillary sign coefficients. Combined with the characteristic symmetry of the two eyeballs, pupillary sign monitoring coefficients are obtained.
It enables quantitative assessment of vital signs in critically ill patients, improving monitoring efficiency and accuracy, and allowing for rapid assessment of patients' health status.
Smart Images

Figure CN118717031B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vital sign data monitoring, analysis and processing, and in particular to a method for monitoring vital signs of critically ill patients based on image processing. Background Art
[0002] Critically ill patients often suffer from multiple organ failure and lack the ability to move independently. Therefore, real-time monitoring of their vital signs is crucial to prevent their condition from worsening and delaying timely rescue, which could reduce their survival rate. Basic vital signs include eight key components: temperature, pulse, respiration, blood pressure, consciousness, pupils, urine output, and skin and mucous membranes. Changes in pupil morphology can help doctors quickly assess a patient's condition and health, providing crucial evidence for further treatment.
[0003] Existing technologies can use image processing technology to collect images of the patient's eyeballs to assist in analyzing and evaluating pupil morphological characteristics, but this is affected by subjective factors and is inefficient. In addition, pupil reflexes also reflect whether the patient's nervous system regulation is abnormal. Therefore, in the process of monitoring the patient's vital signs, it is an urgent problem to quantitatively analyze the patient's vital signs based on the patient's pupil morphology and pupil changes to assist in improving the accuracy and efficiency of vital sign monitoring by relevant medical personnel. Summary of the Invention
[0004] In order to solve the technical problems in the prior art of non-objective quantification of pupil signs and low monitoring efficiency during vital sign monitoring, the present invention aims to provide a method for monitoring vital signs of critically ill patients based on image processing. The technical solution adopted is as follows:
[0005] Obtaining eyeball images of the patient's two eyes at the time to be monitored, and obtaining a sequence of eyeball frame images corresponding to the two eyes within a preset future time period at the time to be monitored, each of the eyeball frame image sequences including all eyeball change images of the corresponding eye after light stimulation; obtaining the pupil and iris in each of the eyeball images and the eyeball change images;
[0006] In each of the eyeball images and the eyeball change image, a pupil offset parameter is obtained according to the relative position distribution of the iris and the pupil; and a corresponding iris reference diameter and pupil reference diameter are obtained according to the shape and size of each of the iris and the pupil;
[0007] In each of the eyeball images, a static pupil sign performance coefficient of each eyeball of the patient is obtained based on the difference between the iris reference diameter and the pupil reference diameter, and the shape characteristics of the iris and the pupil, in combination with the pupil offset parameter; in each of the eyeball frame image sequences, a dynamic pupil sign performance coefficient of each eyeball of the patient is obtained based on the change in the difference between the iris reference diameter and the pupil reference diameter in adjacent eyeball change images and the concentrated characteristics of the pupil offset parameter;
[0008] According to the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of each eyeball, combined with the characteristic symmetry of the two eyeballs, the patient's pupil sign monitoring coefficient is obtained.
[0009] Furthermore, the method for obtaining the pupil offset parameter includes:
[0010] In each of the eyeball images and the eyeball change images, a two-dimensional coordinate system is constructed with the geometric center of the pupil or the iris as the origin, and the pupil offset parameter is obtained based on the coordinate distance between the position coordinates of the geometric center of the iris and the position coordinates of the geometric center of the pupil; the coordinate distance is positively correlated with the pupil offset parameter.
[0011] Furthermore, the method for obtaining the iris reference diameter and the pupil reference diameter includes:
[0012] In each of the eyeball images and the eyeball change image, a plurality of line segments of preset angles are obtained through the corresponding geometric centers of the iris and the pupil, and the length of the line segment intersecting the boundary of the iris is used as the iris diameter, and the length of the line segment intersecting the boundary of the pupil is used as the pupil diameter;
[0013] The iris reference diameter is obtained based on the concentrated features of all the iris diameters at different preset angles; the pupil reference diameter is obtained based on the concentrated features of all the pupil diameters at different preset angles.
[0014] Furthermore, the method for obtaining the static pupil sign performance coefficient includes:
[0015] In each of the eyeball images, according to the difference between the pupil reference diameter and the preset standard pupil diameter and the difference between the iris reference diameter and the preset standard iris diameter, combined with the roundness characteristics of the iris and the pupil, a pupil morphology deviation parameter is obtained;
[0016] A static pupil sign performance coefficient is obtained according to the pupil morphology deviation parameter and the pupil offset parameter; the pupil morphology deviation parameter and the pupil offset parameter are both negatively correlated with the static pupil sign performance coefficient.
[0017] Furthermore, the method for obtaining the pupil morphology deviation parameter includes:
[0018] Obtaining a first morphological shift parameter based on a ratio deviation of a reference diameter ratio of the pupil reference diameter to the iris reference diameter relative to a standard diameter ratio of a preset standard pupil diameter to a preset standard iris diameter, wherein the ratio deviation is positively correlated with the first morphological shift parameter;
[0019] In each of the eyeball images, obtaining a second morphological offset parameter according to the discrete conditions of the pupil diameter and the discrete conditions of the iris diameter at different preset angles;
[0020] A pupil morphology deviation parameter is obtained according to the first morphology deviation parameter and the second morphology deviation parameter, and both the first morphology deviation parameter and the second morphology deviation parameter are positively correlated with the pupil morphology deviation parameter.
[0021] Furthermore, the method for obtaining the second morphological offset parameter includes:
[0022] The variance of the pupil diameter at different preset angles is used as the pupil non-circular parameter; the variance of the iris diameter at different preset angles is used as the iris non-circular parameter; a second morphological offset parameter is obtained based on the pupil non-circular parameter and the iris non-circular parameter, and the pupil non-circular parameter and the iris non-circular parameter are both positively correlated with the second morphological offset parameter.
[0023] Furthermore, the method for obtaining the dynamic pupil sign performance coefficient includes:
[0024] In each eyeball frame image sequence, the accumulated value of the difference between the ratios of the iris reference diameter and the pupil reference diameter in all adjacent eyeball change images is used as a first change parameter;
[0025] The ratio of the preset pupil diameter reduction lower limit to the preset standard iris diameter is used as a dynamic reference ratio; in each eyeball frame image sequence, a second change parameter is obtained based on the concentrated features of the pupil offset parameters in all adjacent eyeball change images and the deviations of all reference diameter ratios relative to the dynamic reference ratio;
[0026] A dynamic pupil sign performance coefficient is obtained according to the first change parameter and the second change parameter, the first change parameter is positively correlated with the dynamic pupil sign performance coefficient, and the second change parameter is negatively correlated with the dynamic pupil sign performance coefficient.
[0027] Furthermore, the method for obtaining the second change parameter includes:
[0028] The total number of reference diameter ratios less than the dynamic reference ratio is used as the deviation number, and the ratio of the deviation number to the total number of all eyeball change images in the eyeball frame image sequence is used as the abnormal dynamic parameter;
[0029] The mean of all pupil shift parameters is taken as the average shift parameter;
[0030] A second change parameter is obtained according to the abnormal dynamic parameter and the average offset parameter, and both the abnormal dynamic parameter and the average offset parameter are positively correlated with the second change parameter.
[0031] Furthermore, the method for obtaining the pupil sign monitoring coefficient includes:
[0032] Merging the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of the same eyeball in a positive correlation manner to obtain a pupil sign performance sub-parameter for each eyeball;
[0033] Merging the pupil sign performance sub-parameters of the patient's two eyes in a positively correlated manner to obtain an initial pupil sign monitoring coefficient;
[0034] Performing negative correlation mapping on the differences in the pupil sign sub-parameters of the patient's two eyes, and using the negative correlation mapping result as a symmetric confidence weight;
[0035] The initial pupil sign monitoring coefficient is weighted using the symmetric confidence weight, and the weighted result is used as the patient's pupil sign monitoring coefficient.
[0036] Furthermore, the preset angles include at least a horizontal direction, a vertical direction, and two diagonal directions.
[0037] The present invention has the following beneficial effects:
[0038] The embodiment of the present invention obtains eyeball images of the patient's two eyeballs at the time to be monitored, and obtains an eyeball frame image sequence corresponding to the two eyeballs within a preset future time period at the time to be monitored, each eyeball frame image sequence includes all eyeball change images of the corresponding eyeball after light stimulation; obtains the pupil and iris in each eyeball image and eyeball change image; obtains the pupil offset parameter in each eyeball image and eyeball change image according to the relative position distribution of the iris and the pupil; obtains the corresponding iris reference diameter and pupil reference diameter according to the shape and size of each iris and pupil; in each eyeball image, obtains the static pupil sign expression of each eyeball of the patient according to the difference between the iris reference diameter and the pupil reference diameter, and the shape characteristics of the iris and the pupil, combined with the pupil offset parameter. The static pupil sign performance coefficient quantitatively analyzes and evaluates the patient's vital sign performance at a static angle where the pupil is not affected by external light. The larger the coefficient, the more normal the vital sign. In each eyeball frame sequence, the dynamic pupil sign performance coefficient of each eyeball of the patient is obtained based on the changes in the differences between the iris reference diameter and the pupil reference diameter in the adjacent eyeball change images, and the concentrated characteristics of the pupil offset parameters. The dynamic pupil sign performance coefficient quantitatively analyzes and evaluates the patient's vital sign performance by analyzing the dynamic expansion and contraction of the pupil. The larger the coefficient, the more normal the vital sign. Based on the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of each eyeball, combined with the characteristic symmetry of the two eyeballs, the patient's pupil sign monitoring coefficient is obtained. Based on image processing technology, the present invention uses iris comparison to evaluate the morphological characteristics and changes of the pupil, and combines the symmetry characteristics of the patient's two eyeballs to quantitatively evaluate the patient's vital signs, thereby improving the efficiency of vital sign monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0040] Figure 1 A flowchart of a method for monitoring vital signs of critically ill patients based on image processing provided by one embodiment of the present invention;
[0041] Figure 2 is a schematic diagram of a pupil and iris according to an embodiment of the present invention;
[0042] Figure 3 A flowchart of a method for obtaining a static pupil sign performance coefficient provided by one embodiment of the present invention;
[0043] Figure 4A flowchart of a method for obtaining a dynamic pupil sign performance coefficient is provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION
[0044] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following, in combination with the accompanying drawings and preferred embodiments, describes in detail the method for monitoring vital signs of critically ill patients based on image processing proposed by the present invention, its specific implementation method, structure, characteristics and effects. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0045] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0046] The following describes in detail a method for monitoring vital signs of critically ill patients based on image processing provided by the present invention with reference to the accompanying drawings.
[0047] See also Figure 1 , which shows a flowchart of a method for monitoring vital signs of critically ill patients based on image processing according to an embodiment of the present invention, specifically comprising:
[0048] Step S1, respectively obtain eye images of the patient's two eyeballs at the time to be monitored, and obtain an eye frame sequence corresponding to the two eyeballs within a preset future time period at the time to be monitored, each eye frame sequence including all eye change images of the corresponding eyeball after light stimulation; obtain the pupil and iris in each eyeball image and the eye change image.
[0049] In one embodiment of the present invention, first, at the time to be monitored, eye images of the left and right eyeballs of a critically ill patient to be monitored are obtained respectively; then a focused flashlight is used to illuminate the middle of the two eyes and then the flashlight is turned off to obtain an eye frame sequence of the patient's two eyeballs within a preset future time period at the time to be monitored, wherein the preset future time period is 3s, that is, the eye frame sequence of each eyeball is collected within 3s at a sampling frequency of 0.1s per frame, and each eyeball contains a total of 30 frames of eye change images; the implementer can use a head-mounted eye tracker or other related instruments to collect eye images and eye frame sequences, and can use an eyelid opener to assist in opening the patient's eyelids when necessary; wherein the pupil and iris can be clearly observed in both the eye image and the eye change image, and all images are further pre-processed such as noise reduction and grayscale conversion.
[0050] It should be noted that the eyeball image is collected under normal indoor lighting conditions without light stimulation, and is mainly used to analyze the pupil morphological characteristics; the eyeball change image is collected after light stimulation, and is mainly used to analyze the pupil changes; the image collection and preprocessing process is a well-known technology and will not be repeated here; the implementer can also set the duration and sampling frequency of the preset future time period by himself.
[0051] To facilitate subsequent analysis of pupil morphological characteristics and changes, an embodiment of the present invention further obtains the pupil and iris in each eyeball image and eyeball change image; in one embodiment of the present invention, a large number of rich eyeball sample images are first collected, the pupil area and iris area in the eyeball sample images are labeled, and then a training set is constructed to train a convolutional neural network model. The collected eyeball images and eyeball change images are input into the trained convolutional neural network model to obtain the pupil and iris in each eyeball image and eyeball change image.
[0052] In other embodiments of the present invention, the implementer may also use threshold segmentation to obtain the iris and pupil based on the color difference of the iris and eyeball, or may obtain them based on an edge detection algorithm. The training and application of the neural network model are all well-known technologies and will not be elaborated here.
[0053] Step S2: In each eyeball image and eyeball change image, pupil offset parameters are obtained according to the relative position distribution of the iris and pupil; and corresponding iris reference diameter and pupil reference diameter are obtained according to the shape and size of each iris and pupil.
[0054] Considering that the iris is a thin, flat, oval membrane located in the middle layer of the eyeball, it can adjust the pupil diameter according to the intensity of external light and control the amount of light entering the eye; the pupil is the circular hole in the center of the iris and is the channel for light to enter the eye; therefore, the embodiment of the present invention first obtains the pupil offset parameter based on the relative position distribution of the iris and pupil; the relative position distribution of the iris and pupil indirectly reflects the patient's physical signs. The closer the pupil is to the center of the iris, the more normal the patient's current physical signs are, and the smaller the pupil offset parameter is.
[0055] It should be noted that, in one embodiment of the present invention, since the eyeball images and eyeball change images are all collected by the same collection device and fixed collection position and angle, the scales of all eyeball images and eyeball change images are the same, and the subsequent calculations of the size and position coordinates of the iris and pupil are not affected by deviations caused by scale; if there is an image scale effect, the implementer can also use the proportional relationship between the pupil and the iris to determine the scale difference for subsequent analysis.
[0056] Preferably, in one embodiment of the present invention, considering that the geometric center can be used to describe the position information of an object, the geometric centers of the pupil and iris are first obtained, and then the relative distance between the geometric centers is used to evaluate the pupil offset; therefore, the method for obtaining the pupil offset parameter includes:
[0057] In each eyeball image and eyeball change image, a two-dimensional coordinate system is constructed with the geometric center of the pupil or iris as the origin, and the pupil offset parameter is obtained based on the coordinate distance between the position coordinates of the geometric center of the iris and the position coordinates of the geometric center of the pupil; the coordinate distance is positively correlated with the pupil offset parameter.
[0058] As an example, a coordinate system is constructed with the geometric center of the pupil as the origin, the position coordinates of the geometric center of the iris are obtained, and the Euclidean distance between the two position coordinates is used as the pupil offset parameter. The larger the Euclidean distance, the greater the deviation between the two geometric centers, and the larger the pupil offset parameter, the lower the possibility that the patient's vital signs are normal. In other examples, a coordinate system can also be constructed based on the geometric center of the iris, and the geometric center can also be replaced by the centroid, center of mass, etc., which will not be repeated here.
[0059] Considering that under normal circumstances, the size and shape of the pupil and iris are usually circular or nearly circular, the embodiment of the present invention further obtains the corresponding iris reference diameter and pupil reference diameter based on the shape and size of each iris and pupil, so as to subsequently analyze the shape and changes of the iris and pupil.
[0060] Preferably, in one embodiment of the present invention, considering that the geometric center can approximately represent the center of the iris and pupil, the length of the line segment passing through the center and intersecting the circle is the diameter; and considering that when the patient's vital signs are abnormal, the iris or pupil may be distorted, the patient's pupil morphology can be accurately assessed based on multi-diameter analysis, avoiding the random coincidence and inaccuracy of single diameter analysis; based on this, the method for obtaining the iris reference diameter and the pupil reference diameter includes:
[0061] In each eyeball image and eyeball change image, a number of line segments with preset angles are obtained through the corresponding geometric centers of the iris and pupil, and the length of the line segment intersecting the boundary of the iris is used as the iris diameter, and the length of the line segment intersecting the boundary of the pupil is used as the pupil diameter;
[0062] The iris reference diameter is obtained based on the concentrated features of all iris diameters at different preset angles; the pupil reference diameter is obtained based on the concentrated features of all pupil diameters at different preset angles.
[0063] The preset angles include at least a horizontal direction, a vertical direction, and two diagonal directions.
[0064] As an example, see Figure 2, which shows a schematic diagram of a pupil and iris in one embodiment of the present invention; Figure 2 The geometric center of the middle pupil coincides with the geometric center of the iris. Draw a horizontal line, a vertical line, and two diagonal lines through the geometric centers. The horizontal line segment 3-7, the vertical line segment 1-5, and the two diagonal line segments 4-8 and 2-6 that intersect with the pupil boundary are used as the pupil diameter; then the mean of all pupil diameters is used as the pupil reference diameter; the iris reference diameter can be obtained in the same way, which is not repeated here.
[0065] It should be noted that when the geometric center of the pupil does not coincide with the geometric center of the iris, the principle of obtaining the iris reference diameter and the pupil reference diameter is the same as when they coincide, and will not be repeated here.
[0066] In other examples, implementers may also set other preset angles to obtain more diameter information for analysis and evaluation; or use the mode or median instead of the mean.
[0067] Step S3, in each eyeball image, based on the difference between the iris reference diameter and the pupil reference diameter, and the shape characteristics of the iris and pupil, combined with the pupil offset parameter, obtain the static pupil sign performance coefficient of each eyeball of the patient; in each eyeball frame sequence, based on the change in the difference between the iris reference diameter and the pupil reference diameter in adjacent eyeball change images, and the concentrated characteristics of the pupil offset parameter, obtain the dynamic pupil sign performance coefficient of each eyeball of the patient.
[0068] Based on the physiological structure of the pupil and iris, the embodiment of the present invention first obtains the static pupil sign performance coefficient of each eyeball of the patient; the static pupil sign performance coefficient is based on the shape and size of the pupil and iris, and combined with the pupil offset. At a static angle where the pupil is not affected by external light, the patient's vital signs at the time to be monitored are quantitatively evaluated. The larger the static pupil sign performance coefficient, the more normal the patient's vital signs at the time to be monitored.
[0069] Preferably, in one embodiment of the present invention, the method for obtaining the static pupil sign performance coefficient includes:
[0070] See also Figure 3 , which shows a flow chart of a method for obtaining a static pupil sign performance coefficient provided by an embodiment of the present invention, specifically comprising:
[0071] Step S301 , in each eyeball image, according to the difference between the pupil reference diameter and the preset standard pupil diameter and the difference between the iris reference diameter and the preset standard iris diameter, combined with the roundness characteristics of the iris and pupil, obtain the pupil morphology deviation parameter.
[0072] Taking into account image acquisition reasons, the size of the iris and pupil in the eye image or the eye change image is not in a one-to-one relationship with the actual size of the iris and pupil of the human body. Directly using the difference between the reference diameter and the preset mark to evaluate the pupil morphological deviation may cause misjudgment; however, the proportional relationship between the pupil reference diameter and the iris reference diameter should have a certain similarity with the proportional relationship between the preset standard pupil diameter and the preset standard iris diameter; therefore, the embodiment of the present invention can use the proportional relationship to indirectly indicate whether the pupil size and the iris size have undergone abnormal changes; and considering that the rounder the shape of the pupil and iris, the more normal it is; based on this, the pupil morphological deviation parameter is obtained.
[0073] In a preferred embodiment of the present invention, the method for obtaining pupil morphology deviation parameters includes:
[0074] Obtaining a first morphological shift parameter according to a ratio deviation of a reference diameter ratio of a pupil reference diameter to a iris reference diameter relative to a standard diameter ratio of a preset standard pupil diameter to a preset standard iris diameter, wherein the ratio deviation is positively correlated with the first morphological shift parameter;
[0075] In each eyeball image, a second morphological offset parameter is obtained according to the discrete conditions of pupil diameter and iris diameter at different preset angles;
[0076] A pupil morphology deviation parameter is obtained according to the first morphology deviation parameter and the second morphology deviation parameter, and both the first morphology deviation parameter and the second morphology deviation parameter are positively correlated with the pupil morphology deviation parameter.
[0077] Among them, the method for obtaining the second morphological offset parameter includes: using the variance of the pupil diameter at different preset angles as the pupil non-circular parameter; using the variance of the iris diameter at different preset angles as the iris non-circular parameter; obtaining the second morphological offset parameter based on the pupil non-circular parameter and the iris non-circular parameter, and the pupil non-circular parameter and the iris non-circular parameter are both positively correlated with the second morphological offset parameter.
[0078] As an example, the calculation formula of the first morphology offset parameter is: ; is the first morphological offset parameter for each eyeball; The reference diameter of the pupil for each eyeball; The reference diameter of the iris for each eyeball; is the reference diameter ratio; The preset standard pupil diameter; It is the preset standard iris diameter; is the standard diameter ratio; is the absolute value symbol; in this example, since the iris diameter standard is 12-14mm and the pupil diameter standard is 2-5mm, we take the middle value respectively, then Take 3.5, Take 13, then the standard diameter ratio after common denominator is ; The smaller the ratio deviation, the smaller the first morphological offset parameter, which means the more normal the vital signs.
[0079] As an example, the calculation formula of the second morphology offset parameter is: ; The second morphological offset parameter for each eyeball; is the iris non-circularity parameter of each eyeball; is the pupil non-circular parameter of each eyeball; in this example, the pupil non-circular parameter and the iris non-circular parameter are added and combined to obtain the second morphological offset parameter. The implementer can also use positive correlation merging methods such as multiplication or weighted summation, which will not be repeated here; the greater the difference in iris diameter under different preset angles, the larger the corresponding iris non-circular parameter; similarly, the same is true for the pupil non-circular parameter; the more the pupil and iris shapes are similar to circles, the smaller the corresponding non-circular parameter, and the smaller the second morphological offset parameter.
[0080] In another example, the shape characteristics of the pupil and iris can be evaluated based on the roundness formula to obtain the second morphological offset parameter; the roundness formula is: ,in is the area, is the circumference; after calculating the roundness, perform the inverse operation to obtain the corresponding pupil non-circular parameters and iris non-circular parameters, and then obtain the second morphological offset parameters, which will not be repeated here.
[0081] In one example of the present invention, after obtaining the first morphological offset parameter and the second morphological offset parameter, the product of the two is used as the pupil morphological deviation parameter; in other embodiments of the present invention, the implementer can also perform weighted merging or correlation mapping on his own to amplify the sensitivity to one of them, which will not be repeated here.
[0082] Step S302: Obtain a static pupil sign performance coefficient based on the pupil morphology deviation parameter and the pupil offset parameter; both the pupil morphology deviation parameter and the pupil offset parameter are negatively correlated with the static pupil sign performance coefficient.
[0083] As an example, the pupil morphology deviation parameter and the pupil offset parameter are multiplied and combined, and then negative correlation mapping is performed to obtain the static pupil sign performance coefficient, such as performing a reciprocal operation, which is as follows: ,in The static pupil sign performance coefficient of each eyeball; The pupil offset parameter for each eye; is the pupil morphology deviation parameter for each eyeball; 0.01 is an anti-zero adjustment parameter to ensure that the denominator is meaningful. Implementers can set other non-zero positive parameters according to actual needs; in other examples, implementers can also use the product of the two as an exponential function with the natural constant e as the base in , and negative correlation mapping can be performed; static pupil sign performance coefficients can also be obtained based on other basic mathematical operations and related mapping methods, such as or etc., no further elaboration.
[0084] At this point, the static pupil sign performance coefficient of each eyeball of the patient is obtained.
[0085] Taking into account that the iris adjusts the pupil diameter according to the intensity of external light, it is a necessary sign to evaluate whether the patient's autonomic nervous system and vital signs are normal; and considering that the eye frame sequence records a series of eye change images of the patient after light stimulation, analyzing the dynamic changes of the pupil can help evaluate the patient's pupil expansion and contraction ability, and then evaluate the patient's vital signs; and light does not affect the changes in the position distribution of the patient's pupil; therefore, the embodiment of the present invention further obtains the dynamic pupil sign performance coefficient of each eye of the patient in each eye frame sequence according to the changes in the difference between the iris reference diameter and the pupil reference diameter in adjacent eye change images, and the concentrated characteristics of the pupil offset parameters; the larger the dynamic pupil sign performance coefficient, the more normal the patient's vital signs.
[0086] Preferably, in one embodiment of the present invention, the method for obtaining the dynamic pupil sign performance coefficient includes:
[0087] See also Figure 4 , which shows a flow chart of a method for obtaining a dynamic pupil sign performance coefficient provided by one embodiment of the present invention, specifically comprising:
[0088] Step S401 : In each eyeball frame sequence, the accumulated difference between the ratios of the iris reference diameter to the pupil reference diameter in all adjacent eyeball change images is used as a first change parameter.
[0089] When the patient's nervous system can quickly respond to light within a very short period of time while the images of adjacent eyeballs change, causing pupil changes, it indicates that the stronger the pupil's ability to expand and contract, the more normal the patient's vital signs are. At the same time, considering that the size of the iris does not change under light stimulation, using the iris as a reference can more accurately assess pupil changes.
[0090] Therefore, as an example, in each eyeball frame sequence, the ratio of the iris reference diameter to the pupil reference diameter in each eyeball change image is obtained; then the absolute value of the difference between the ratios in two adjacent eyeball change images is calculated, and the accumulated value of the absolute value of the difference is directly used as the first change parameter.
[0091] It should be noted that, without being affected by the image scale, the implementer may also directly use the accumulated sum of the absolute values of the differences in pupil reference diameters in all adjacent eyeball change images as the first change parameter.
[0092] In step S402, the ratio of the preset lower limit of pupil diameter reduction to the preset standard iris diameter is used as a dynamic reference ratio; in each eye frame sequence, the second change parameter is obtained based on the concentrated characteristics of the pupil offset parameters in all adjacent eye change images and the deviation of all reference diameter ratios relative to the dynamic reference ratio.
[0093] Considering that pupil contraction and dilation also have certain limitations, when the limit range is exceeded, it can be determined that the pupil is over-constricted or over-dilated, which also reflects that the patient's vital signs are in an unhealthy state; further combined with the pupil deviation, the vital sign-related parameters are evaluated to obtain a second change parameter; the second change parameter reflects whether the pupil changes excessively and has positional deviation. The larger the second change parameter, the more abnormal the patient's vital signs.
[0094] Therefore, in a preferred embodiment of the present invention, considering that pupils primarily constrict first and then gradually return to normal under light stimulation, the proportion of pupils with excessive constriction can be evaluated in all eye change images. A greater proportion indicates that excessive constriction may have occurred. At the same time, the mean can reflect the concentrated characteristics of the data. Therefore, the method for obtaining the second change parameter includes:
[0095] The total number of reference diameter ratios smaller than the dynamic reference ratio is taken as the number of deviations, and the ratio of the number of deviations to the total number of all eyeball change images in the eyeball frame sequence is taken as the abnormal dynamic parameter;
[0096] The mean of all pupil shift parameters is taken as the average shift parameter;
[0097] A second change parameter is obtained according to the abnormal dynamic parameter and the average offset parameter. Both the abnormal dynamic parameter and the average offset parameter are positively correlated with the second change parameter.
[0098] As an example, the calculation formula of the second variation parameter is: ;in, is a second variation parameter for each eye frame sequence; is the number of deviations in each eye frame sequence; The total number of all eye change images in each eye frame sequence; is the abnormal dynamic parameter of each eye frame sequence; The average offset parameter for each eye frame sequence.
[0099] In this example, the method for obtaining the deviation amount is: set the preset pupil diameter reduction lower limit to 1.5mm, and then divide the ratio of the preset pupil diameter reduction lower limit to the preset standard iris diameter by the common denominator to obtain the dynamic reference ratio , and then the number of deviations is obtained, which is not described in detail here. In other examples, the implementer may also combine the abnormal dynamic parameters and the average offset parameters by using positive correlation methods such as addition or weighted summation, or may use variance to evaluate the concentrated characteristics of the pupil offset parameters, which is not described in detail here.
[0100] Step S403 , obtaining a dynamic pupil sign performance coefficient according to the first change parameter and the second change parameter, wherein the first change parameter is positively correlated with the dynamic pupil sign performance coefficient, and the second change parameter is negatively correlated with the dynamic pupil sign performance coefficient.
[0101] As an example, the second variation parameter is taken as an exponential function with the natural constant e as the base. in , perform negative correlation mapping, and then multiply and combine the negative correlation mapping result with the first change parameter to obtain the dynamic pupil sign performance coefficient; in other examples, the implementer may also adopt other negative correlation mapping methods, which will not be repeated here.
[0102] At this point, the dynamic pupil sign performance coefficient of each eyeball is obtained.
[0103] Step S4, obtaining the patient's pupil sign monitoring coefficient based on the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of each eyeball and the characteristic symmetry of the two eyeballs.
[0104] Since the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient evaluate and quantify the patient's vital sign health status from different angles; and since the vital signs reflected by the patient's two eyeballs should be similar or consistent and have certain symmetrical characteristics; therefore, the embodiment of the present invention further obtains the patient's pupil sign monitoring coefficient based on the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of each eyeball, combined with the characteristic symmetry of the two eyeballs.
[0105] Preferably, in one embodiment of the present invention, the pupil sign performance sub-parameters of each eyeball are first obtained by combining the two angles; considering that the two eyeballs have a certain characteristic symmetry, the larger the pupil sign performance sub-parameters of the two eyeballs are, the healthier the patient's vital signs are; but if the difference between the pupil sign performance sub-parameters of the two eyeballs is, the lower the confidence level of the pupil sign performance sub-parameters of the two eyeballs is or the more abnormal the patient's vital signs are; based on this, the method for obtaining the pupil sign monitoring coefficient includes:
[0106] The static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of the same eyeball are positively correlated and combined to obtain the pupil sign performance sub-parameter of each eyeball;
[0107] The pupil sign sub-parameters of the patient's two eyes were positively correlated and combined to obtain the initial pupil sign monitoring coefficient;
[0108] Perform negative correlation mapping on the differences in pupillary sign sub-parameters of the patient's two eyes, and use the negative correlation mapping results as symmetric confidence weights;
[0109] The initial pupil sign monitoring coefficient is weighted using symmetric confidence weights, and the weighted result is used as the patient's pupil sign monitoring coefficient.
[0110] As an example, the pupil sign monitoring coefficient is calculated as: ;in is the patient's pupillary sign monitoring coefficient; is the static pupil sign performance coefficient of the left eye; is the dynamic pupil sign performance coefficient of the left eye; is the static pupil sign performance coefficient of the right eye; is the dynamic pupil sign performance coefficient of the right eye; is the initial pupil sign monitoring coefficient; is the absolute value symbol; is the initial pupil sign monitoring coefficient; is the symmetric confidence weight.
[0111] In this example, the implementer multiplies and combines the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of the same eye, and then multiplies and combines the pupil sign performance sub-parameters of different eyeballs; the absolute value of the difference between the pupil sign performance sub-parameters of different eyeballs is added with the anti-zero parameter 0.01 and then the inverse operation is performed to perform negative correlation mapping. The anti-zero parameter implementer can adjust it at will. In other examples, the implementer can also use positive correlation merging methods such as addition or weighting, or use the absolute value of the difference between the pupil sign performance sub-parameters of different eyeballs as an exponential function with the natural constant e as the base. in , perform negative correlation mapping to obtain symmetric confidence weights; I will not go into details here.
[0112] At this point, the patient's pupillary sign monitoring coefficient has been quantified. A higher pupillary sign monitoring coefficient indicates a more normal patient's vital signs. Implementers can further combine this with seven other vital sign indicators to assist medical staff in comprehensively assessing the patient's current vital signs and developing the next step in physical therapy.
[0113] In summary, the embodiments of the present invention respectively obtain eyeball images and eyeball frame image sequences of the patient's two eyeballs; further obtain the pupil and iris in the image; further obtain the pupil offset parameter based on the relative position distribution of the iris and pupil, and obtain the iris reference diameter and pupil reference diameter in each image; obtain the static pupil sign performance coefficient of each eyeball of the patient in each eyeball image; obtain the dynamic pupil sign performance coefficient of each eyeball of the patient in each eyeball frame image sequence; and then obtain the patient's pupil sign monitoring coefficient in combination with the characteristic symmetry of the two eyeballs. Based on image processing technology, the present invention uses iris comparison to evaluate the morphological characteristics and changes of the pupil, and combines the symmetry characteristics of the patient's two eyeballs to quantitatively evaluate the patient's vital signs and improve the efficiency of vital sign monitoring.
[0114] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0115] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.
Claims
1. A method for monitoring vital signs of critically ill patients based on image processing, characterized in that: The method comprises: Obtain eyeball images of the patient's two eyeballs at the time to be monitored, and obtain eyeball frame image sequences corresponding to the two eyeballs within a preset future time period at the time to be monitored, each of the eyeball frame image sequences including all eyeball change images of the corresponding eyeball after light stimulation; obtain the pupil and iris in each of the eyeball images and the eyeball change images; In each of the eyeball images and the eyeball change images, pupil displacement parameters are obtained according to the relative position distribution of the iris and the pupil; and corresponding iris reference diameters and pupil reference diameters are respectively obtained according to the shape and size of each of the iris and the pupil; In each of the eyeball images, according to the difference between the iris reference diameter and the pupil reference diameter, and the shape characteristics of the iris and the pupil, combined with the pupil offset parameter, the static pupil sign performance coefficient of each eyeball of the patient is obtained; in each of the eyeball frame image sequences, according to the change of the difference between the iris reference diameter and the pupil reference diameter in adjacent eyeball change images, and the concentrated characteristics of the pupil offset parameter, the dynamic pupil sign performance coefficient of each eyeball of the patient is obtained; According to the static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of each eyeball, combined with the characteristic symmetry of the two eyeballs, the patient's pupil sign monitoring coefficient is obtained; The method for obtaining the pupil shift parameter includes: In each of the eyeball images and the eyeball change images, a two-dimensional coordinate system is constructed with the geometric center of the pupil or the iris as the origin, and a pupil shift parameter is obtained according to the coordinate distance between the position coordinates of the geometric center of the iris and the position coordinates of the geometric center of the pupil; the coordinate distance is positively correlated with the pupil shift parameter; The method for obtaining the iris reference diameter and the pupil reference diameter includes: In each of the eyeball images and the eyeball change images, a number of line segments of preset angles are obtained through the corresponding geometric centers of the iris and the pupil, and the length of the line segment intersecting the boundary of the iris is used as the iris diameter, and the length of the line segment intersecting the boundary of the pupil is used as the pupil diameter; According to the concentrated features of all the iris diameters at different preset angles, an iris reference diameter is obtained; according to the concentrated features of all the pupil diameters at different preset angles, a pupil reference diameter is obtained; The method for obtaining the static pupil sign performance coefficient includes: In each of the eyeball images, according to the difference between the pupil reference diameter and the preset standard pupil diameter and the difference between the iris reference diameter and the preset standard iris diameter, combined with the roundness characteristics of the iris and the pupil, a pupil morphology deviation parameter is obtained; According to the pupil morphology deviation parameter and the pupil offset parameter, a static pupil sign performance coefficient is obtained; the pupil morphology deviation parameter and the pupil offset parameter are both negatively correlated with the static pupil sign performance coefficient; The method for obtaining the dynamic pupil sign performance coefficient includes: In each eyeball frame image sequence, the difference accumulation value of the ratio of the iris reference diameter to the pupil reference diameter in all adjacent eyeball change images is used as the first change parameter; The ratio of the preset pupil diameter reduction lower limit to the preset standard iris diameter is used as a dynamic reference ratio; in each eyeball frame image sequence, the second change parameter is obtained according to the concentrated features of the pupil offset parameters in all adjacent eyeball change images and the deviations of all reference diameter ratios relative to the dynamic reference ratio; Acquire a dynamic pupil sign performance coefficient according to the first change parameter and the second change parameter, wherein the first change parameter is positively correlated with the dynamic pupil sign performance coefficient, and the second change parameter is negatively correlated with the dynamic pupil sign performance coefficient; The method for obtaining the pupil sign monitoring coefficient includes: The static pupil sign performance coefficient and the dynamic pupil sign performance coefficient of the same eyeball are positively correlated and combined to obtain the pupil sign performance sub-parameter of each eyeball; The pupil sign performance sub-parameters of the two eyeballs of the patient are positively correlated and combined to obtain an initial pupil sign monitoring coefficient; Performing negative correlation mapping on the differences of the pupil sign expression sub-parameters of the two eyeballs of the patient, and using the negative correlation mapping result as the symmetric confidence weight; The initial pupil sign monitoring coefficient is weighted using the symmetric confidence weight, and the weighted result is used as the patient's pupil sign monitoring coefficient.
2. According to the method for monitoring vital signs of critically ill patients based on image processing as described in claim 1, it is characterized in that: The method for obtaining the pupil morphology deviation parameter includes: Obtaining a first morphological shift parameter according to a reference diameter ratio of the pupil reference diameter to the iris reference diameter relative to a ratio deviation of a standard diameter ratio of a preset standard pupil diameter to a preset standard iris diameter, wherein the ratio deviation is positively correlated with the first morphological shift parameter; In each of the eyeball images, a second morphological shift parameter is obtained according to the discrete conditions of the pupil diameter and the discrete conditions of the iris diameter at different preset angles; A pupil morphology deviation parameter is obtained according to the first morphology deviation parameter and the second morphology deviation parameter, and both the first morphology deviation parameter and the second morphology deviation parameter are positively correlated with the pupil morphology deviation parameter.
3. According to the method for monitoring vital signs of critically ill patients based on image processing as described in claim 2, it is characterized in that: The method for obtaining the second form offset parameter includes: The variance of the pupil diameter at different preset angles is used as a pupil non-circular parameter; the variance of the iris diameter at different preset angles is used as an iris non-circular parameter; a second morphological offset parameter is obtained based on the pupil non-circular parameter and the iris non-circular parameter, and both the pupil non-circular parameter and the iris non-circular parameter are positively correlated with the second morphological offset parameter.
4. According to the method for monitoring vital signs of critically ill patients based on image processing as described in claim 1, it is characterized in that: The method for obtaining the second change parameter includes: The total number of reference diameter ratios less than the dynamic reference ratio is taken as the deviation number, and the ratio of the deviation number to the total number of all eyeball change images in the eyeball frame image sequence is taken as the abnormal dynamic parameter; The mean of all the pupil shift parameters is taken as the average shift parameter; A second variation parameter is obtained according to the abnormal dynamic parameter and the average deviation parameter, and both the abnormal dynamic parameter and the average deviation parameter are positively correlated with the second variation parameter.
5. A method for monitoring vital signs of critically ill patients based on image processing according to claim 1, characterized in that: The preset angles at least include a horizontal direction, a vertical direction and two diagonal directions.
Citation Information
Patent Citations
Controlling a luminous means having at least two semiconductor light sources
CN105744881A
Device and system for monitoring an eye of a subject
CN107920733A
System and method for optical assessment of pupillary psychological sensory responses
CN115334957A
Pupil monitoring and evaluating system for general anesthesia patient
CN115813352A
Auxiliary eye nursing intelligent control system based on data analysis
CN116999017A
