Skin moisture analysis method and system for premature infants

Abstract

The present invention discloses a method for analyzing skin moisture of premature infants, including: using an inter-frame difference method to perform facial tracking on the facial area of ​​an infant to collect a facial video of the infant; extracting the ROI area of ​​the infant in each frame of the video; segmenting the infant ROI area and performing three-primary color separation on the segmented image to generate an RGB three-channel image, and processing the skin pixels of each channel image to generate pre-processed digital signals of three channels of green, blue B and red; standardizing the original signal to generate three new signals; then performing an FFT transformation to obtain a spectrum diagram, and selecting the maximum peak frequency as a pulse rate estimation value; according to the obtained pulse rate estimation value, performing a correlation analysis based on the green channel signal as the source signal to obtain the absolute value of the correlation coefficient, and the change in the correlation coefficient is the change in skin moisture. The present invention has great application potential, and has the advantages of high accuracy in synchronous measurement of heart rate, strong anti-noise interference ability, and great application potential.

Description

Technical Field

[0001] The present disclosure relates to the field of big data technology, and in particular to a premature infant skin moisture analysis method and system thereof. Background Art

[0002] Due to the immaturity of regulatory functions, premature infants have unstable body fluids and are prone to imbalance. This is a characteristic of their body fluid regulation. Therefore, it is particularly important to regulate and manage water and electrolytes for premature infants. For extremely low birth weight children, the changes in water and sodium in the early stage are closely related to physiological weight loss, and water and electrolyte disorders are more likely to occur, which requires close monitoring. Factors that affect individual demand include daily loss, including explicit and implicit water loss, dynamic weight changes, etc. It is particularly important to replenish water and electrolytes according to individual needs and to manage infusions well.

[0003] The function of the skin temperature regulation center is not perfect, the subcutaneous fat is thin, the limbs are often stretched, and the body surface area is relatively large, which increases the heat dissipation. The incubator is the main insulation equipment for premature infants. The moderate temperature is the environmental temperature required to maintain normal body temperature with the lowest oxygen consumption according to the age and weight of premature infants. Exceeding this temperature range of ±2°C will affect the metabolism and body temperature of the baby. The moderate temperature of premature infants is different from that of full-term newborns. The younger the gestational age, the higher the moderate temperature, and the moderate temperature gradually decreases with the increase of age. Therefore, premature infants placed in incubators need to adjust the moderate temperature according to their weight and age every day. It is particularly important to adjust the temperature and humidity of the incubator. In clinical practice, the box temperature is adjusted by pre-adjusting the skin temperature of premature infants, but it is rare to evaluate the comprehensive situation of the water content and temperature response of the child through the humidity of the skin. Summary of the invention

[0004] The present disclosure aims at one or more of the above-mentioned existing problems and proposes a method for analyzing skin moisture of premature infants.

[0005] According to one aspect of the present disclosure, a method for analyzing skin moisture of premature infants is provided, comprising the following steps: S100: first, using an inter-frame difference method to perform facial tracking on the facial region of the infant, and then using a video acquisition module to capture a video of the infant's face;

[0006] S200: extracting the ROI region of the baby in each frame of the video according to the baby's facial video through an image preprocessing module, and identifying skin pixels from the pixels in the ROI;

[0007] S300: Segment the ROI area of ​​the infant and perform three-primary color separation on the segmented image to generate an RGB three-channel image, and average the skin pixels of each channel image, and use the average value as the feature value of the frame image in the corresponding color channel to generate pre-processed digital signals of the green, blue and red three channels;

[0008] S400: The preprocessed original signal x R (t), x G (t) and x B (t) Perform standardization processing to obtain a standardized signal, and use the standardized signal as the input signal to perform blind source separation to generate three new signals. The separated source signal is recorded as y a (t), y b (t) and y c (t);

[0009] S500: performing bandpass filtering signal processing on the separated source signal, and performing frequency spectrum analysis on the filtered signal, that is, performing FFT transformation on the signal to obtain a spectrum diagram, and selecting the maximum peak frequency as the pulse rate estimation value in combination with a sliding window algorithm;

[0010] S600: According to the acquired pulse rate estimation value, a correlation analysis is performed based on the green channel signal as the source signal to obtain the absolute value of the correlation coefficient. The change of the correlation coefficient is the change of skin moisture.

[0011] In some implementations, in step S100, face tracking is performed on the video image using a frame difference method, including the following steps:

[0012] S101: Record the nth frame and the n-1th frame image in the video sequence as f n and f n-1 ;

[0013] S102: Obtain the grayscale values ​​of corresponding pixels in the two frames of images and record them as f n (x,y) and f n-1 (x,y);

[0014] S103: The grayscale values ​​of the corresponding pixels in the two frames of images are calculated. n (x,y) and f n-1 (x, y) are subtracted to obtain the corresponding difference, and the absolute value of the difference is taken to obtain the difference value Dn, D n (x,y)=f n (x,y)-f n-1 (x,y)

[0015] S104: Set a threshold value T, compare the pixels with the threshold value T one by one, and perform binarization processing to obtain a binary image formula Rn', which is specifically:

[0016]

[0017] Among them, when Dn exceeds a certain threshold, it is determined to be a moving target point, the point with a gray value of 255 is the moving target point, and the point with a gray value of 0 is the background point;

[0018] S105: Perform connectivity analysis on the acquired Rn', and finally obtain an image Rn containing a complete moving target.

[0019] In some implementations, in step S101, identifying the skin pixel includes:

[0020] Position each ROI grid and confirm the skin pixels within each ROI grid;

[0021] Get the average pixel intensity and number of skin pixels in each ROI grid.

[0022] In some implementations, in step S101, extracting the ROI region of the baby in each frame of the video includes:

[0023] Acquire multiple frames of target images including facial images of infants to be detected;

[0024] Performing image preprocessing on the multiple frames of target images to obtain key point information corresponding to each frame of the image;

[0025] Based on the key point information corresponding to each frame of image, an identification area of ​​the region of interest corresponding to each frame of image is determined.

[0026] In some embodiments, in step S102, the pre-processed original signal is subjected to standardization processing to obtain a standardized signal, comprising the following steps:

[0027] Assume that the original signal is a real random variable with mean zero;

[0028] The following formula is used to perform zero mean processing on the original signal;

[0029] x' i (t) = (x i (t)-μ i ) / σ i

[0030] where i = 1, 2, 3, μ i and σ i are the mean and standard deviation of x(t) respectively. After the standardized transformation of x(t), the mean of x'(t) is zero and the variance is 1.

[0031] In some implementations, in step S105, a correlation analysis is performed based on the acquired pulse rate estimate using the green channel signal as the source signal, specifically including:

[0032] According to the extracted skin pixels, the pulse rate-related characteristic signals are extracted according to the time dimension;

[0033] Selecting sample points from the extracted pulse rate-related characteristic signals, and performing signal fusion according to the selected sample points;

[0034] The pulse rate is calculated based on the fused pulse rate related characteristic signals to obtain a continuously generated pulse rate time series.

[0035] Another invention further provides a premature infant skin moisture analysis system, which adopts the premature infant skin moisture analysis method, comprising:

[0036] A video image acquisition module, used for acquiring video images of infants;

[0037] A face tracking module, used to perform face tracking on the video image of the baby and obtain an image of the baby's facial area;

[0038] An image preprocessing module is used to extract the ROI area of ​​the baby in each frame of the video based on the baby's facial video;

[0039] A segmentation processing module is used to perform image segmentation processing on the ROI area image of the infant and convert the segmented image signal into a three-channel digital signal;

[0040] The blind source separation module is used to standardize the standard deviation of the generated three-channel digital signal and separate it into three independent signals y a (t), y b (t) and y c (t);

[0041] The pulse rate extraction module is used to extract the three independent signals y a (t), y b (t), y c (t) performing bandpass filtering signal processing and performing frequency spectrum analysis on the filtered signal, i.e., performing FFT transformation on the filtered signal to obtain a spectrum diagram, and selecting the maximum peak frequency as the pulse rate estimation value in combination with a sliding window algorithm;

[0042] The correlation processing module performs correlation analysis based on the acquired pulse rate estimation value and the green channel signal as the source signal to obtain the absolute value of the correlation coefficient. The change of the correlation coefficient is the change of skin moisture.

[0043] In some embodiments, a lighting correction module is also included for performing lighting compensation on the baby's facial video captured by the camera using a histogram equalization method.

[0044] The beneficial effects of the present invention are: by using the inter-frame difference method to perform facial tracking on the facial area of ​​the baby to obtain the baby's facial video; extracting the ROI area of ​​the baby in each frame of the video according to the baby's facial video, identifying the skin pixels from the pixels within the ROI; performing correlation analysis on the pulse rate estimate obtained by processing the skin pixels based on the green channel signal as the source signal, obtaining the absolute value of the correlation coefficient, and the change in the correlation coefficient is the change in skin moisture. Therefore, the detection efficiency is improved, the detection operation is convenient and safe, and there is no need to contact the baby, which solves the inconvenience of baby detection and ensures the safety of the baby. The present invention has great application potential, and has the advantages of high measurement accuracy, strong anti-noise interference ability, and great application potential. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in describing the embodiments of the present application are briefly introduced below.

[0046] Figure 1 is a flow chart of the method for analyzing skin moisture in premature infants disclosed in the present invention;

[0047] Figure 2 A waveform diagram of the standardized processing of the first channel of the premature infant skin moisture analysis method disclosed in the present invention;

[0048] Figure 3 A waveform diagram of the standardized processing of the second channel of the premature infant skin moisture analysis method disclosed in the present invention;

[0049] Figure 4 Schematic diagram of the standardized processing waveform of the third channel of the premature infant skin moisture analysis method disclosed in the present invention;

[0050] Figure 5 Schematic diagram of plotting curves of the premature infant skin moisture analysis method disclosed in the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0052] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0053] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0054] The present disclosure will be further described in detail below in conjunction with the accompanying drawings.

[0055] Blind Source Separation (BSS) is a signal processing method that estimates the source signal and the unknown mixed channel parameters only by observing the sensor signal when the source signal and the mixed channel parameters are unknown; Independent Component Analysis (ICA); ROI (region of interest). In machine vision and image processing, the area to be processed is outlined in the image to be processed in the form of a box, circle, ellipse, irregular polygon, etc., which is called the region of interest, or ROI.

[0056] Embodiment 1:

[0057] Reference Manual Attached Figure 1 -5, showing a method for analyzing skin moisture of premature infants, comprising the following steps:

[0058] S1OO: First, the facial region of the infant is tracked by using an inter-frame difference method, and then the facial video of the infant is collected by a video acquisition module;

[0059] S1O1: extracting the ROI region of the baby in each frame of the video according to the baby's facial video through the image preprocessing module, and identifying skin pixels from the pixels in the ROI;

[0060] S1O2: Segment the ROI area of ​​the infant and perform three-primary color separation on the segmented image to generate an RGB three-channel image, and take the average value of the skin pixels in each channel image, and use the average value as the feature value of the frame image in the corresponding color channel to generate pre-processed digital signals of the green, blue and red three channels;

[0061] S1O3: The preprocessed original signal x R (t), x G (t) and x B (t) Perform standardization processing to obtain a standardized signal, and use the standardized signal as the input signal to perform blind source separation to generate three new signals. The separated source signal is recorded as y a (t), y b (t) and y c (t);

[0062] S1O4: perform bandpass filtering on the separated source signal, and perform frequency spectrum analysis on the filtered signal, that is, perform FFT transformation on it to obtain a spectrum diagram, and select the maximum peak frequency as the pulse rate estimation value in combination with the sliding window algorithm;

[0063] S1O5: Based on the acquired pulse rate estimate, a correlation analysis is performed based on the green channel signal as the source signal to obtain the absolute value of the correlation coefficient. The change in the correlation coefficient is the change in skin moisture.

[0064] Specifically, a correlation curve can be drawn to display intuitively. In this embodiment, a correlation analysis method is used to perform correlation analysis on the three independent signals after blind source separation and the green channel signal in the source signal, that is, to study the similarity of the two signals and find the most relevant independent signal as the required signal, and then the signal screening is completed.

[0065] like Figure 5 As shown, in Matlab, the corrcoef function is used to calculate and find the ICA signal corresponding to the correlation coefficient with the largest absolute value, which is the selected signal; in Matlab, the crosscorr function is used to implement the correlation analysis using a set of collected data as an example. The absolute values ​​of the correlation coefficients are 0.9715, 0.0592 and 0.2293, respectively. Therefore, the first ICA signal is most correlated with the green channel source signal. These three values ​​are the values ​​related to the skin tissue we need, and the change of this value is the skin tissue change rate we need. We mainly use the green channel source signal to calculate the ratio of the green channel to the red channel, the green channel to the blue channel, and the red channel to the blue channel. The change in the tissue structure of skin moisture also changes the correlation of its RGB signals, thus realizing the monitoring of infant skin moisture.

[0066] In this embodiment, in step S100, face tracking is performed on the video image using the frame difference method, including the following steps:

[0067] S101: Record the nth frame and the n-1th frame image in the video sequence as f n and fn-1 ;

[0068] S102: Obtain the grayscale values ​​of corresponding pixels in the two frames of images and record them as f n (x,y) and f n-1 (x,y);

[0069] S103: The grayscale values ​​of the corresponding pixels in the two frames of images are calculated. n (x,y) and f n-1 (x, y) are subtracted to obtain the corresponding difference, and the absolute value of the difference is taken to obtain the difference value Dn, D n (x,y)=f n (x,y)-f n-1 (x,y)

[0070] S104: Set a threshold value T, compare the pixels with the threshold value T one by one, and perform binarization processing to obtain a binary image formula Rn', which is specifically:

[0071]

[0072] Among them, when Dn exceeds a certain threshold, it is determined to be a moving target point, the point with a gray value of 255 is the moving target point, and the point with a gray value of 0 is the background point;

[0073] S105: Perform connectivity analysis on the acquired Rn', and finally obtain an image Rn containing a complete moving target.

[0074] In this embodiment, in step S101, identifying the skin pixels includes:

[0075] Position each ROI grid and confirm the skin pixels within each ROI grid;

[0076] Get the average pixel intensity and number of skin pixels in each ROI grid.

[0077] In this embodiment, in step S101, extracting the ROI region of the baby in each frame of the video includes:

[0078] Acquire multiple frames of target images including facial images of infants to be detected;

[0079] Performing image preprocessing on the multiple frames of target images to obtain key point information corresponding to each frame of the image;

[0080] Based on the key point information corresponding to each frame of image, an identification area of ​​the region of interest corresponding to each frame of image is determined.

[0081] In this embodiment, in step S102, the pre-processed original signal is subjected to standardization processing to obtain a standardized signal, which includes the following steps:

[0082] Assume that the original signal is a real random variable with mean zero;

[0083] The following formula is used to perform zero mean processing on the original signal;

[0084] x' i (t) = (x i (t)-μ i ) / σ i

[0085] where i = 1, 2, 3, μ i and σ i are the mean and standard deviation of x(t) respectively. After the standardized transformation of x(t), the mean of x'(t) is zero and the variance is 1.

[0086] In this embodiment, in step S105, according to the acquired pulse rate estimation value, a correlation analysis is performed based on the green channel signal as the source signal, specifically including:

[0087] According to the extracted skin pixels, the pulse rate-related characteristic signals are extracted according to the time dimension;

[0088] Selecting sample points from the extracted pulse rate-related characteristic signals, and performing signal fusion according to the selected sample points;

[0089] The pulse rate is calculated based on the fused pulse rate related characteristic signals to obtain a continuously generated pulse rate time series.

[0090] Another invention further provides a premature infant skin moisture analysis system, which adopts the premature infant skin moisture analysis method, comprising:

[0091] A video image acquisition module, used for acquiring video images of infants;

[0092] A face tracking module, used to perform face tracking on the video image of the baby and obtain an image of the baby's facial area;

[0093] An image preprocessing module is used to extract the ROI area of ​​the baby in each frame of the video based on the baby's facial video;

[0094] A segmentation processing module is used to perform image segmentation processing on the ROI area image of the infant and convert the segmented image signal into a three-channel digital signal;

[0095] The blind source separation module is used to standardize the standard deviation of the generated three-channel digital signal and separate it into three independent signals y a(t), y b (t) and y c (t);

[0096] The pulse rate extraction module is used to extract the three independent signals y a (t), y b (t), y c (t) performing bandpass filtering signal processing and performing frequency spectrum analysis on the filtered signal, i.e., performing FFT transformation on the filtered signal to obtain a spectrum diagram, and selecting the maximum peak frequency as the pulse rate estimation value in combination with a sliding window algorithm;

[0097] The correlation processing module performs correlation analysis based on the acquired pulse rate estimation value and the green channel signal as the source signal to obtain the absolute value of the correlation coefficient. The change of the correlation coefficient is the change of skin moisture.

[0098] Specifically, it also includes a lighting correction module for using a histogram equalization method to perform lighting compensation on the baby's facial video captured by the camera.

[0099] The beneficial effects of the present invention are as follows: the facial area of ​​the baby is tracked by using the inter-frame difference method to obtain the baby's facial video; the ROI area of ​​the baby in each frame of the video is extracted based on the baby's facial video, and the skin pixels are identified from the pixels within the ROI; the pulse rate estimate obtained by processing the skin pixels is analyzed based on the green channel signal as the source signal to obtain the absolute value of the correlation coefficient, and the change in the correlation coefficient is the change in skin moisture. Therefore, the detection efficiency is improved, the detection operation is convenient and safe, and there is no need to contact the baby, which solves the inconvenience of baby detection and ensures the safety of the baby. The present invention has great application potential, and has the advantages of high measurement accuracy, strong anti-noise interference ability, and great application potential.

[0100] It should be noted that the system provided in the above embodiment, when realizing its functions, only uses the division of the above functional modules as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiment belong to the same concept, and their specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0101] The above are only some embodiments of the present disclosure. For those skilled in the art, several modifications and improvements can be made without departing from the creative concept of the present disclosure, and these all belong to the protection scope of the present disclosure.

Claims

1. A method for analyzing skin moisture in premature infants, characterized in that: The following steps are involved: S100: Using the frame difference method to track the facial area of ​​the baby, and then using the video acquisition module to capture the baby's facial video; S200: extracting the ROI region of the baby in each frame of the video according to the baby's facial video through an image preprocessing module, and identifying skin pixels from the pixels in the ROI; S300: Segment the ROI area of ​​the infant and perform three-primary color separation on the segmented image to generate an RGB three-channel image, and average the skin pixels of each channel image, and use the average value as the feature value of the frame image in the corresponding color channel to generate pre-processed digital signals of the green, blue and red three channels; S400: The preprocessed original signal x R (t), x G (t) and x B (t) Perform standardization processing to obtain a standardized signal, and use the standardized signal as the input signal to perform blind source separation to generate three new signals. The separated source signal is recorded as y a (t), y b (t) and y c (t); S5 00: Perform bandpass filtering on the separated source signal, perform correlation analysis based on the green channel signal as the source signal, and obtain the absolute value of the correlation coefficient. The change in the correlation coefficient is the change in skin moisture, where the correlation coefficient includes the ratio of the green channel to the red channel, the ratio of the green channel to the blue channel, and the ratio of the red channel to the blue channel.

2. The method for analyzing skin moisture of premature infants according to claim 1, characterized in that: In step S100, face tracking is performed on the video image using the frame difference method, including the following steps: Record the nth frame and n-1th frame image in the video sequence as f n and f n-1 ; Get the grayscale values ​​of the corresponding pixels in the two frames of images and record them as f n (x,y) and f n-1 (x,y); The gray value f of the corresponding pixel in the two frames of image n (x,y) and f n-1 (x, y) are subtracted to obtain the corresponding difference, and the absolute value of the difference is taken to obtain the difference value D n , D n (x,y)=f n (x,y)-f n-1 (x, y) Set the threshold T, compare the pixels with the threshold T one by one for binarization, and get the binary image formula R n ＇ , specifically: Among them, when D n When it exceeds a certain threshold, it is determined as a moving target point. The point with a gray value of 255 is the moving target point, and the point with a gray value of 0 is the background point. S105: The obtained R n ＇ Connectivity analysis is performed, and finally the image R containing the complete moving target can be obtained. n .

3. The method for analyzing skin moisture of premature infants according to claim 1, characterized in that: In step S200, identifying the skin pixels includes: Position each ROI grid and confirm the skin pixels within each ROI grid; Get the average pixel intensity and number of skin pixels in each ROI grid.

4. The method for analyzing skin moisture of premature infants according to claim 2, characterized in that: In step S200, the ROI region of the baby in each frame of the video is extracted, including: Acquire multiple frames of target images including facial images of infants to be detected; Performing image preprocessing on the multiple frames of target images to obtain key point information corresponding to each frame of the image; Based on the key point information corresponding to each frame of image, an identification area of ​​the region of interest corresponding to each frame of image is determined.

5. The method for analyzing skin moisture of premature infants according to claim 1, characterized in that: In step S300, the pre-processed original signal is subjected to standardization processing to obtain a standardized signal, which includes the following steps: Assume that the original signal is a real random variable with mean zero; The following formula is used to perform zero mean processing on the original signal; x ＇ i (t)=(x i (t)-m i ) / s i where i = 1, 2, 3, μ i and σ i are the mean and standard deviation of x(t), respectively. ＇ (t) has a mean of zero and a variance of 1.

6. The method for analyzing skin moisture of premature infants according to claim 1, characterized in that: In step S500, three independent signals y a (t), y b (t), y c (t) performing bandpass filtering signal processing and performing frequency spectrum analysis on the filtered signal, i.e., performing FFT transformation on it to obtain a spectrum diagram, selecting the maximum peak frequency as the pulse rate estimation value in combination with a sliding window algorithm, and performing correlation analysis based on the green channel signal as the source signal according to the obtained pulse rate estimation value, i.e., studying the similarity of the two signals and obtaining the correlation coefficient.

7. Premature infant skin moisture analysis system, characterized in that: The method for analyzing skin moisture of premature infants according to any one of claims 1 to 6 comprises: A video image acquisition module, used for acquiring video images of infants; A face tracking module, used to perform face tracking on the video image of the baby and obtain an image of the baby's facial area; An image preprocessing module is used to extract the ROI area of ​​the baby in each frame of the video based on the baby's facial video; A segmentation processing module is used to perform image segmentation processing on the ROI area image of the infant and convert the segmented image signal into a three-channel digital signal; The blind source separation module is used to standardize the standard deviation of the generated three-channel digital signal and separate it into three independent signals y a (t), y b (t) and y c (t); The pulse rate extraction module is used to extract the three independent signals y a (t), y b (t), y c (t) performing bandpass filtering signal processing and performing frequency spectrum analysis on the filtered signal, i.e., performing FFT transformation on the filtered signal to obtain a spectrum diagram, and selecting the maximum peak frequency as the pulse rate estimation value in combination with a sliding window algorithm; The correlation processing module performs correlation analysis based on the acquired pulse rate estimation value and the green channel signal as the source signal to obtain the absolute value of the correlation coefficient. The change of the correlation coefficient is the change of skin moisture.

8. The premature infant skin moisture analysis system according to claim 7, characterized in that: It also includes a lighting correction module, which is used to use a histogram equalization method to perform lighting compensation on the baby's facial video captured by the camera.

Images