A standardized pulse diagnosis method and platform based on a wearable pulse instrument

Abstract

The application provides a standardized pulse diagnosis method based on a wearable pulse instrument, which comprises the following steps: initialization, acquisition of wrist pulse pressure data by multiple pressurizations from low to high, curve fitting, determination of a reference wave, identification of a wave trough, calculation of a single mean wave, multi-dimensional evaluation of pulse position, pulse number, pulse shape and pulse potential, etc., collection of human pulse by the wearable pulse instrument, combination of a corresponding pulse processing method, simulation of the principle of pulse wave change felt by a doctor by finger pressing in traditional Chinese medicine, simulation of artificial pressing pressure by a wristband airbag, monitoring and collection of pulse change data after pressurization by a sensor, generation of a human pulse change curve by an AI algorithm after conversion of electric signals and pressure values, and substitution of subjective judgment of artificial pulse pressing. The accuracy of artificial pulse pressing is affected by the mood of the pulse pressing person and external environmental factors, while the sensor can continuously and objectively monitor the pulse change of a patient, so that more objective pulse diagnosis data can be obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traditional Chinese medicine pulse diagnosis, and particularly relates to a standardized pulse diagnosis method based on a wearable pulse instrument. BACKGROUND

[0002] Pulse diagnosis has the advantages of non-invasiveness, simplicity, rapidness, and comprehensive reflection of overall function state, and plays an irreplaceable role in the traditional Chinese medicine syndrome differentiation and treatment system. However, pulse diagnosis is difficult to learn, master and popularize, so the modernization research of traditional Chinese medicine pulse diagnosis, and the realization of data, visualization and standardization are of great significance to the development of traditional Chinese medicine.

[0003] Objective pulse diagnosis: pulse diagnosis is one of the unique diagnostic methods of traditional Chinese medicine. It mainly uses the feeling of fingers to analyze the characteristics of pulse wave such as "position, number, shape and potential", so as to judge the function state of viscera and achieve the purpose of non-invasive diagnosis, which has a positive significance for the diagnosis and treatment of diseases. However, traditional Chinese medicine pulse diagnosis has a certain subjectivity, and it is quite difficult to accurately master and use. It is said that "it is easy in the heart, but difficult to understand under the finger". This project can improve the objectivity of traditional Chinese medicine pulse diagnosis.

[0004] Visual pulse diagnosis: traditional Chinese medicine pulse diagnosis is a text description, which is relatively abstract. The diagnosis of pulse is entirely based on personal experience and understanding in clinic. Although images are used in modern times, they are mostly schematic diagrams, pattern diagrams and model diagrams, which are not real pulse diagrams. There is a big difference in nature between the actual pulse diagram and the measured pulse diagram. This project will realize the visualization of traditional Chinese medicine pulse diagnosis and change the current situation that traditional Chinese medicine pulse diagnosis is not suitable for popularization and difficult to learn.

[0005] Standardized pulse diagnosis: the standard of traditional pulse diagnosis is not unified. With the progress of science and technology, the research of pulse diagnosis instrument and the emergence of artificial intelligence make it possible to realize the standardization of pulse diagnosis. However, many units are enthusiastic about researching pulse instruments in the laboratory, and rarely consider the clinical application problems, which leads to the disconnection between research results and clinic. The pulse instruments developed are either complicated to operate, inconvenient to use, or too expensive for clinical units to adopt. It is difficult to analyze the results of pulse examination, and there is a lack of professional guidance. Most traditional Chinese medicine doctors do not understand the meaning of pulse diagram, so it is difficult to popularize and promote in clinic. Most importantly, most of the collected pulse waves are not analyzed, there are images without analysis, or different methods, each to its own, and there is no reverse verification, which makes it difficult to form the standardization of machine pulse diagnosis. This project is expected to realize the standardization of pulse diagnosis.

[0006] In order to realize the standardization and objectivity of traditional Chinese medicine pulse diagnosis, the applicant invents a standardization pulse diagnosis method based on a wearable pulse instrument, collects human pulse through the wearable pulse instrument, combines the corresponding pulse processing method, simulates the principle of pulse wave change by finger pressing pulse in traditional Chinese medicine, uses wristband airbag pressure simulation artificial pressing pressure, sensor monitors and collects pulse change data after pressure, AI algorithm completes the conversion of electric signal and pressure value, and generates human pulse change curve, so as to replace the subjective judgment of artificial pulse taking. The accuracy of artificial pulse taking is influenced by the mood of pulse taking person and external environmental factors, while the sensor can continuously and objectively monitor the pulse change of the patient, and then more objective pulse diagnosis data is obtained. SUMMARY

[0007] The application provides a standardization pulse diagnosis method based on a wearable pulse instrument, and solves the problems of objectivity and standardization of traditional Chinese medicine pulse diagnosis in the prior art.

[0008] The technical scheme of the application is as follows: a standardization pulse diagnosis method based on a wearable pulse instrument, using a wearable pulse instrument to collect pulse, comprising the following steps:

[0009] Step S1: initializing the height h and weight g of the patient, and calculating the standard pressure P0

[0010] Step S2: according to the following pulse cutting methods under the standard pressure, the pressure is sequentially pressed: pressure floating, pressure medium, pressure sinking, and pressure pressing the bone, and the above each step T seconds, m pressure data are read every second, the data of the first T0 seconds of each step are removed, and then fitting is performed, and the fitting curves are sequentially marked as ②: floating fitting curve; ③: medium fitting curve; ④: sinking fitting curve; and ⑤: pressing the bone fitting curve;

[0011] Step S3: determining the reference wave and calculating the average wavelength; taking the fitting curve with the maximum minimum convex hull area Ca as the reference wave by calculating the minimum convex hull area of each fitting curve; filtering, setting the wavelength or wave height that is too small as a pseudo-wave, and ignoring (or merging into an adjacent wave with smaller wavelength) the wave trough and the adjacent smaller wave peak of the pseudo-wave. Recalculate the average wavelength as aL.

[0012] Step S4: wave trough identification;

[0013] The point in the fitted curve where the slope changes is defined as the pole. The segment with the largest difference in the vertical axis between adjacent concave and convex poles is the ascending segment. The starting point of the ascending segment is the calibration trough point. Using aL as the scale, after the calibration trough point, find the ascending segments whose starting points are within the range of [aL×constant F] to [aL×(1+constant F^2)] on the X-axis. If no ascending segment is found within this range, the range is modified to [aL×(1+constant F)] to [aL×(2+constant F^2)]. If no ascending segment is found, it is reported as an invalid spectrum. Then, the starting point of this ascending segment is used as the next calibration point, and so on, to determine the starting points of all ascending segments after the calibration trough point. Similarly, before the calibration trough point, find the rising segment whose starting point is within the range of [-aL×constant F] to [-aL×(1+constant F^2)] on the X-axis. If no rising segment is found within this range, adjust the range to [-aL×(1+constant F)] to [-aL×(2+constant F^2)]. If no rising segment is found, report it as an invalid spectrum, and use this rising segment's starting point as the next calibration point. Continue this process to determine the starting points of all rising segments before the calibration trough point. The lowest concave pole within the range of [-aL×constant F^3 to 0] before and after each rising segment's starting point is the preceding trough. The highest point between preceding and following troughs is the wave crest.

[0014] Step S5: Calculate the single mean wave;

[0015] Calculate the single-wave average: Divide each single wave along the x-axis into N equal parts to obtain x0 to x... N The N+1 points correspond to y0~y N , y0~y N Take the average value of each, and combine these N+1 average values ​​into a single wave; where the wave area is the area of ​​the figure enclosed by the line connecting the pre-peak valley and the post-peak valley and the fitted curve; the wave circumference is the sum of the arc length and chord length of the figure enclosed by the line connecting the pre-peak valley and the post-peak valley / chord and the fitted curve / arc.

[0016] Step S6: Based on the waveform obtained in the above steps, calculate the pulse position, pulse number, pulse shape, and pulse strength according to each dimension.

[0017] As a preferred technical solution, the standard pressure P0 is calculated using the following formula:

[0018] P0 = 90 × (BMI / 23)^0.5

[0019] Wherein, BMI is the Body Mass Index, and BMI = g / h2, where height h is in meters and weight g is in kilograms. P0×(constant G)^0 (floating pressure); P0×(constant G)^1 (central pressure); P0×(constant G)^2 (sinking pressure); P0×(constant G)^3 (pressing pressure on the bone).

[0020] As a preferred technical solution, the pulse position is calculated according to the depth, length, frequency, rhythm, width, tension, fullness, fluency, force and slope in step S6.

[0021] As a preferred technical solution, the pulse position is calculated from the depth and length, and the single uniform wave height is divided into shallow pulse position, deep pulse position and extremely deep pulse position by comparing the single uniform wave height between the second fitting curve, the third fitting curve, the fourth fitting curve and the fifth fitting curve of the current single uniform wave; the single uniform wave height is divided into short pulse position and long pulse position by comparing the ratio of the part above the wave waist of the wave area to the part below the wave waist of the wave area between the reference wave of the current single uniform wave and the system average value.

[0022] As a preferred technical solution, the pulse number is calculated from the frequency and rhythm, and the frequency of the pulse number is divided into slow, slightly slow, fast and extremely fast by comparing the 15-second wave peak count between the reference wave of the current single uniform wave and the system average value; the rhythm is divided into uneven and pause by comparing the wave height or wave length within the reference wave curve of the current single uniform wave.

[0023] As a preferred technical solution, the pulse shape is calculated from the width, tension and fullness, and the width is divided into wide and thin according to the ratio of the single uniform wave area to the 15-second wave peak count between the reference wave of the current single uniform wave and the system average value; the tension is divided into tight and soft according to the wave peak y-axis value between the reference wave of the current single uniform wave and the system average value; the fullness is divided into empty and full according to the wave valley y-axis value between the multiple single uniform waves.

[0024] As a preferred technical solution, the pulse shape is calculated from the width, tension and fullness, and the width is divided into wide and thin according to the ratio of the single uniform wave area to the 15-second wave peak count between the reference wave of the current single uniform wave and the system average value; the tension is divided into tight and soft according to the wave peak y-axis value between the reference wave of the current single uniform wave and the system average value; the fullness is divided into empty and full according to the wave valley y-axis value between the multiple single uniform waves.

[0025] A standardized pulse diagnosis platform based on a wearable pulse instrument is used to perform the standardized pulse diagnosis method based on the wearable pulse instrument.

[0026] Compared with the prior art, the present application has the following beneficial effects:

[0027] A standardized pulse diagnosis method based on a wearable pulse instrument is provided, which collects human pulse through the wearable pulse instrument, combines corresponding pulse processing methods, simulates the principle of pulse wave change felt by a TCM finger pressing pulse, adopts a wristband airbag to simulate artificial pressing pressure, a sensor monitors and collects pulse change data after pressure, an AI algorithm converts electrical signals into pressure values, and generates human pulse change curves to replace subjective judgment of artificial pulse taking. BRIEF DESCRIPTION OF DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0029] Figure 1 A method flowchart of the standardized pulse diagnosis method based on the wearable pulse instrument of the present application;

[0030] Figure 2 A pulse position, pulse number, pulse shape and pulse potential calculation method of the standardized pulse diagnosis method based on the wearable pulse instrument of the present application;

[0031] Figure 3 A first interface screenshot of the standardized pulse diagnosis platform based on the wearable pulse instrument of the present application;

[0032] Figure 4 A second interface screenshot of the standardized pulse diagnosis platform based on the wearable pulse instrument of the present application;

[0033] Figure 5 A third interface screenshot of the standardized pulse diagnosis platform based on the wearable pulse instrument of the present application;

[0034] Figure 6 A fourth interface screenshot of the standardized pulse diagnosis platform based on the wearable pulse instrument of the present application;

[0035] Figure 7 A fifth interface screenshot of the standardized pulse diagnosis platform based on the wearable pulse instrument of the present application. DETAILED DESCRIPTION

[0036] The technical solutions of the present application will be described clearly and completely below in connection with the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative efforts belong to the scope of the present application.

[0037] With reference to Figure 1 The embodiment provides a standardized pulse diagnosis method based on a wearable pulse instrument, and specifically comprises the following steps.

[0038] Step 1: initialize height (m) and weight (Kg), calculate standard pressure = 90*(BMI / 23)^0.5, and BMI (body mass index) = weight / height^2;

[0039] Step 2: float according to the pressure of [standard pressure], take the pressure of [standard pressure] * (constant G)^1, sink according to the pressure of [standard pressure * (constant G)^2], and press according to the pressure of [standard pressure * (constant G)^3] along the bone, sequentially press, read 25 pressure data per second according to 20 seconds per step, eliminate the data in the first 5 seconds of each step, and sequentially mark the fitting curves as ② curve, ③ curve, ④ curve and ⑤ curve; if any one of ②, ③, ④ and ⑤ is missing, the calculation is ended, and an invalid spectrum is fed back.

[0040] Step 3, reference wave: calculate the minimum convex hull area Ca of the fitting curve, If any one of ②, ③, ④ and ⑤ is (highest point-lowest point) / (Ca / 15 seconds)>constant G^6, the calculation is ended, and an invalid spectrum is fed back. Filtering: if the average wavelength of the initial calculation is pL, and the average wave height is pH, the wavelength <pL*constant F and the wave height <pH*constant F are set as pseudo waves, and the wave troughs and adjacent smaller wave peaks of the pseudo waves are marked and ignored (or merged into adjacent smaller wavelengths). Recalculate the average wavelength aL. The X-axis projection of the front wave trough is B, the X-axis projection of the rear wave trough is C, the X-axis projection of the front wave trough of the previous wave is A, if the pseudo wave is the first wave, remove the front wave trough, if the pseudo wave is the tail wave, remove the rear wave trough, if |(B-A)-pL|<|(C-A)-pL|, remove the rear wave trough, if |(B-A)-pL|>|(C-A)-pL|, remove the front wave trough, and if |(B-A)-pL|=|(C-A)-pL|, remove the wave trough with a larger y.

[0041] Fourth step, trough identification: there are some extreme points in the curve (i.e. points where the slope changes ±), and the adjacent concave-convex extreme points are Δy = max ∩ Δy > pH × constant F^3, which is the ascending section. Compare the starting point of the ascending section with the second extreme point before the starting point (the first extreme point is convex), and the lower one is the calibration trough point. Take the newly calculated average wavelength aL as the ruler, find the starting point of the ascending section in the X-axis [aL × constant F] ~ [aL × (1 + constant F^2)] range after the calibration trough point, if no ascending section is found in this range, modify the range to [aL × (1 + constant F)] ~ [aL × (2 + constant F^2)], if no ascending section is still found, feedback as invalid spectrum, and take the starting point of the ascending section as the next calibration point, and so on to determine the starting point of all ascending sections after the calibration trough point. Similarly, find the starting point of the ascending section in the X-axis [-aL × constant F] ~ [-aL × (1 + constant F^2)] range before the calibration trough point, if no ascending section is found in this range, modify the range to [-aL × (1 + constant F)] ~ [-aL × (2 + constant F^2)], if no ascending section is still found, feedback as invalid spectrum, and take the starting point of the ascending section as the next calibration point, and so on to determine the starting point of all ascending sections before the calibration trough point. If the first ascending section Δy < pH × constant F, the corresponding trough point needs to be removed. The lowest concave extreme point in the [-aL × constant F^3 ~ 0] range before and after each ascending section starting point is the front trough. The highest point between the front and rear troughs is the wave peak.

[0042] Fifth step, calculate single average wave: divide each single wave x-axis 24 into equal parts, get x0-x24, 25 equally divided points corresponding to y0-y24, take the average of y0-y24 respectively, and combine the 25 average values into a single wave ② ③ ④ ⑤; Wave area: the area of the figure enclosed by the connecting line of the peak before and after the trough and the fitting curve; Wave perimeter: the perimeter (arc length + chord length) of the figure enclosed by the connecting line (chord) of the peak before and after the trough and the fitting curve (arc).

[0043] Sixth step, calculate pulse position, pulse number, pulse shape, and pulse potential according to depth, length, frequency, rhythm, width, tension, fullness, fluency, force, and slope.

[0044] The specific calculation method of the sixth step is referred to in Figure 2 Specifically, the pulse position is calculated from two dimensions of depth and length, and the single average wave height between the second fitting curve, the third fitting curve, the fourth fitting curve and the fifth fitting curve is compared to divide it into shallow pulse position, deep pulse position and extremely deep pulse position; the ratio of the part above the wave waist to the part below the wave waist of the wave area is compared between the reference wave of the single average wave and the system average value to divide it into short pulse position and long pulse position.

[0045] The pulse number is calculated from two dimensions of frequency and rhythm. The frequency of the pulse number is divided into slow, slightly slow, fast and very fast by comparing the 15-second peak count of the reference wave of the single average wave with the system average value. The rhythm is divided into uneven and pause by comparing the wave height or length within the reference wave curve of the single average wave.

[0046] The pulse shape is calculated from three dimensions of width, tension and fullness. The width is divided into wide and thin according to the ratio of the single average wave area to the 15-second peak count by comparing the reference wave of the single average wave with the system average value. The tension is divided into tight and soft according to the peak y-axis value by comparing the reference wave of the single average wave with the system average value. The fullness is divided into empty and full according to the trough y-axis value by comparing multiple single average waves.

[0047] The pulse potential is calculated from three dimensions of fluency, strength and slope. The fluency is divided into smooth and harsh by comparing the reference wave of the single average wave with the system average value. The strength is divided into powerless and powerful according to the ratio of the single average wave area to the single average wave length by comparing the reference wave of the single average wave with the system average value. The slope is divided into slow and fast according to the ratio of the peak x-axis value to the length by comparing the reference wave within the single average wave with the system average value.

[0048] Meanwhile, the applicant also develops a pulse diagnosis instrument management system matched only, such as Figures 3 to 7 as shown

[0049] Figure 3 : boot page;

[0050] Figure 4 : no height weight data page: if no data is received after booting, prompt the user account and fill in the height weight information in the client;

[0051] Figure 5 : height weight data page: if data is received after booting, the accepted data is displayed, and the user is prompted to confirm whether the device is bound, and click

start

[0052] Figure 6 : collecting pulse page: in turn, float, take, sink, and press the bone four kinds of pressure, and the yellow arrow at the bottom of the text flashes when the corresponding step is reached;

[0053] Figure 7 : complete: display the collection is completed, and prompt the user to view the results in the client.

[0054] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A standardized pulse diagnosis method based on a wearable pulse diagnostic instrument, characterized in that, Includes the following steps: Step S1: Initialize the patient's height h and weight g, and calculate the standard pressure P0; Step S2: Apply pressure sequentially according to the following standard pulse diagnosis techniques: pressure on the surface, pressure on the middle, pressure on the deep, and pressure on the bone. Each pressure application lasts for T seconds, and m pressure data points are read per second. After discarding the data from the first T0 seconds of each step, the data is fitted. The fitted curves are labeled as follows: ② Surface fitting curve; ③ Middle fitting curve; ④ Deep fitting curve; ⑤ Pressure on the bone fitting curve. Step S3: Determine the reference wave and calculate the average wavelength; By calculating the minimum convex hull area Ca of each fitted curve, the fitted curve with the largest minimum convex hull area is taken as the reference wave; filtering is performed, and those with too small wavelength or wave height are set as pseudo waves. The trough of the pseudo wave and a nearby smaller wave peak are marked as ignored or merged into an adjacent wave with a smaller wavelength, and the average wavelength is recalculated as aL. Step S4: Valley identification; The point in the fitted curve where the slope changes is defined as the pole. The segment with the largest difference in the vertical axis between adjacent concave and convex poles is the rising segment. The starting point of the rising segment is the calibration trough point. Using aL as the scale, after the calibration trough point, find the rising segments whose starting points are in the range of [aL×constant F] to [aL×(1+constant F^2)] on the X-axis. If no rising segment is found in this range, the range is modified to [aL×(1+constant F)] to [aL×(2+constant F^2)]. If no rising segment is found, it is reported as an invalid spectrum. Then, the starting point of this rising segment is used as the next calibration point, and so on, to determine the starting points of all rising segments after the calibration trough point. Similarly, before the calibration trough point, find the rising segment whose starting point is in the range of [-aL×constant F]~[-aL×(1+constant F^2)] on the X-axis. If no rising segment is found in this range, the range is modified to [-aL×(1+constant F)]~[-aL×(2+constant F^2)]. If no rising segment is found, it is reported as an invalid spectrum. Then, the starting point of this rising segment is used as the next calibration point. This process is repeated to determine the starting points of all rising segments before the calibration trough point. The lowest concave pole in the range of [-aL×constant F^3~0] before and after each rising segment starting point is the preceding trough, and the highest point between the preceding and following troughs is the wave crest. Step S5: Calculate the single mean wave; Calculate the single-wave average: Divide each single wave along the x-axis into N equal parts to obtain x0~x N The N+1 evenly distributed points correspond to y0~y N , y0~y N Take the average value of each, and combine these N+1 average values ​​into a single wave; where the wave area is the area of ​​the figure enclosed by the line connecting the pre-peak valley and the post-peak valley and the fitted curve; the wave circumference is the sum of the arc length and chord length of the figure enclosed by the line connecting the pre-peak valley and the post-peak valley / chord and the fitted curve / arc. Step S6: Based on the waveform obtained in the above steps, calculate the pulse position, pulse number, pulse shape, and pulse momentum according to 10 dimensions: depth, length, frequency, rhythm, width, tension, fullness, fluency, strength, and slope.

2. The standardized pulse diagnosis method based on a wearable pulse diagnostic instrument as described in claim 1, characterized in that, The formula for calculating the standard pressure P0 is as follows: P0 = 90 × (BMI / 23)^0.5 Wherein, BMI is the Body Mass Index, and BMI=g / h2, where height h is in meters and weight g is in kilograms. The pressure is applied sequentially according to the following steps: pressure buoyancy P0×(constant G)^0, pressure meandering P0×(constant G)^1, pressure sinking P0×(constant G)^2, and pressure heavy pressing P0×(constant G)^3 onto the bones.

3. The standardized pulse diagnosis method based on a wearable pulse diagnostic instrument as described in claim 1, characterized in that, in, The pulse position is calculated from two dimensions: depth and length. Based on the comparison of the single average wave height among the second, third, fourth, and fifth fitting curves of this single average wave, the pulse position is divided into shallow pulse position, deep pulse position, and extremely deep pulse position. Based on the comparison of the reference wave of this single average wave with the system average value, the pulse position is divided into short pulse position and long pulse position by comparing the ratio of the portion above the wave waist to the portion below the wave waist of the wave area.

4. The standardized pulse diagnosis method based on a wearable pulse diagnostic instrument as described in claim 1, characterized in that, in, Pulse count is calculated from two dimensions: frequency and rhythm. The frequency of the pulse count is divided into slow, slightly slow, fast, and very fast by comparing the reference wave of this single average wave with the system average wave and counting the peaks over 15 seconds. The rhythm is divided into irregular and paused by comparing the wave height or wavelength within the reference wave curve.

5. A standardized pulse diagnosis method based on a wearable pulse diagnostic instrument as described in claim 1, characterized in that, in, The pulse shape is calculated from three dimensions: width, tension, and fullness. Width is divided into wide and narrow based on the ratio of the single average wave area to the 15-second peak count, compared with the system average wave. Tension is divided into tight and loose based on the peak y-axis value, compared with the system average wave. Fullness is divided into tense and empty based on the trough y-axis value, compared with multiple single average waves.

6. The standardized pulse diagnosis method based on a wearable pulse diagnostic instrument as described in claim 1, characterized in that, The pulse strength is calculated from three dimensions: smoothness, intensity, and slope. The smoothness is divided into smooth and rough by comparing the single-average wave reference wave with the system average value. The intensity is divided into weak and strong by comparing the ratio of the single-average wave area to the single-average wave perimeter. The slope is divided into gentle and rapid by comparing the ratio of the wave crest x-axis value to the wavelength within the single-average wave reference wave.

7. A standardized pulse diagnosis platform based on a wearable pulse diagnostic instrument, characterized in that, This is used to perform a standardized pulse diagnosis method based on a wearable pulse diagnostic instrument as described in any one of claims 1 to 6.

Images