Diagnosis assisting program

Abstract

A diagnostic support program capable of displaying organ motion is provided. The diagnostic support program analyzes images of a human organ and displays the analysis result, and causes a computer to execute: a process of acquiring a plurality of frame images, a process of calculating periodic changes in the state of the organ between the frame images; a process of performing Fourier transform on the periodic changes in the state of the organ, a process of extracting a frequency spectrum in a fixed frequency band including a frequency spectrum corresponding to the frequency of the organ motion from the frequency spectrum obtained after the Fourier transform, a process of performing inverse Fourier transform on the frequency spectrum extracted by the fixed frequency band, and a process of outputting each image after the inverse Fourier transform.

Description

Technical Field

[0001] This invention relates to a technique for analyzing cardiac images and displaying the analysis results. Background Technology

[0002] In recent years, the rising mortality rate from heart disease has increased the necessity for practical and simple diagnostic techniques. MRI diagnostic technology is rapidly developing, and its importance in the imaging diagnosis of the cardiac region is constantly increasing due to its ability to perform various cardiac examinations in a short time. Regarding "cardiac MRI," examinations such as "cinnamic MRI," "perfusion," "delayed contrast enhancement," and "black blood BB" are performed. Specifically, compared to ultrasound and SPECT, cine MRI, without limitations in its examination range, can observe any cross-section and has high reproducibility. Therefore, it is commonly used in many medical institutions. For example, in cine MRI, data acquisition of approximately 10 slices / 20 phases of the entire left ventricle is performed using an electrocardiogram-gated method. Recently, high contrast between blood and myocardium has been achieved using the "steady-state method." Furthermore, in "cardiac function analysis," the need for evaluating cardiac function via MRI is increasing because it provides more accurate values ​​compared to CT, LVG (left ventriculography), and SPECT.

[0003] As mentioned above, cardiac MRI is clinically useful, and specifically, imaging is performed in cine MRI at many medical institutions, but software analysis of the images is rarely used. This is reportedly because software used for traditional cardiac function analysis requires complex operations to extract and correct the contours of the medial and lateral myocardium. Furthermore, the reproducibility of the analysis results is also problematic, as the tracking of the medial and lateral myocardial contours is easily influenced by the operator. In addition, with the widespread use of cardiac MRI and the increasing number of medical institutions using MRI systems from multiple manufacturers, the different sequence names used by each company make it difficult to easily process this data.

[0004] To address the aforementioned issues and reduce the workload of tracing complex myocardial contours, software has been developed that improves accuracy and automatically performs interpolation even during unintentional tracking, thereby reducing correction work. This software also reduces the stress of image observation by displaying images side-by-side in a browser for MRI cardiac function analysis and click-based operations.

[0005] Existing technical documents

[0006] Non-patent documents

[0007] Non-patent literature 1:

[0008] https: / / www.zio.co.jp / ziostation2 / Summary of the Invention

[0009] The problem to be solved by the present invention

[0010] However, as described in Non-Patent Document 1, simply displaying MRI images side-by-side makes it difficult for doctors to grasp the pathological condition. Therefore, it is desirable to display images based on the state of the heart. That is, it is desirable to control the human heart as an object and display images representing actual movement based on the waveform or frequency change trend of the heart or its images.

[0011] The present invention was made in view of this situation, and its object is to provide a diagnostic support program capable of displaying organ movement. More specifically, the object is to generate an image for auxiliary diagnosis by calculating numerical values ​​for auxiliary diagnosis, by digitizing the consistency rate or other inconsistency rate of the waveform and Hz of the new target data to be measured, and further by imaging these numerical values.

[0012] Problem Solving Methods

[0013] (1) To achieve the above objectives, this application takes the following steps. That is, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results. The program causes a computer to perform the following process, which includes: processing to acquire multiple frame images; processing to calculate a periodic change in organ state between each frame image; processing to perform a Fourier transform on the periodic change in organ state; processing to extract the spectrum obtained after Fourier transform from the spectrum in a fixed frequency band containing the spectrum corresponding to the organ motion frequency; processing to perform an inverse Fourier transform on the spectrum extracted from the fixed frequency band; and processing to output each image after performing the inverse Fourier transform.

[0014] (2) Further, the diagnostic support program according to one aspect of the present invention further includes: processing of segmenting an image of a human organ into multiple block regions and calculating the image changes in each block region of each frame image; processing of performing a Fourier transform on the image changes in each block region of each frame image; processing of extracting the spectrum from a fixed frequency band containing the spectrum corresponding to the frequency of organ movement to obtain the spectrum after Fourier transform; and processing of performing an inverse Fourier transform on the spectrum extracted from the fixed frequency band.

[0015] (3) Further, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following process, the process comprising: processing to acquire multiple frame images; processing to calculate a periodic change in organ state between each frame image; processing to calculate a periodic change in organ state between each frame image; processing to extract pixels that vary with the frequency corresponding to the periodic change in organ state in the image by means of a digital filter; and processing to output an image comprising the pixels extracted by the digital filter.

[0016] (4) Further, the diagnostic support program according to one aspect of the present invention further includes: processing of segmenting an image of an organ into multiple block regions and calculating the changes in the image in each block region of each frame image; and processing of extracting pixels corresponding to the changes in the frequency of organ movement in the image by means of a digital filter.

[0017] (5) Further, according to one aspect of the present invention, the diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following process, the process including: processing to acquire multiple frame images; processing to calculate the periodic rate of change of organ state between each frame image; processing to select a color corresponding to the periodic rate of change of organ state; and processing to add the selected color to the rate of change of pixel value and display the image on the display.

[0018] (6) Further, the diagnostic support program according to one aspect of the present invention further includes: processing of segmenting an image of an organ into multiple block regions and calculating the rate of change of the image in each block region of each frame image; and processing of selecting a color corresponding to the rate of change of the pixel value in each block region.

[0019] (7) Further, according to one aspect of the present invention, the diagnostic support program is a diagnostic support program for analyzing human images and displaying the analysis results, which causes a computer to perform the following process, which includes: processing for acquiring multiple frame images; processing for defining an analysis range for all acquired frame images; processing for dividing the analysis range into multiple regions using the von Ronoite ssellation method; and processing for performing any arithmetic operations to be performed on periodic changes on each segmented region.

[0020] (8) Further, according to one aspect of the present invention, the diagnostic support program is a diagnostic support program for analyzing human images and displaying analysis results, the program causing a computer to perform the following process, the process including: processing for acquiring multiple frame images; processing for defining an analysis range for all acquired frame images; processing for dividing the analysis range into multiple regions; and processing for classifying each region based on an index of periodic changes in each segmented region.

[0021] (9) Further, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results. The program causes a computer to perform the following process, which includes: processing to acquire multiple frame images; processing to calculate periodic changes characterizing the organ while maintaining the absolute positional relationship of each pixel in the multiple frame images; processing to perform a Fourier transform on the periodic changes characterizing the organ state; processing to extract the spectrum obtained after Fourier transform from the spectrum in a fixed frequency band including the spectrum corresponding to the frequency of organ movement; processing to perform an inverse Fourier transform on the spectrum extracted from the fixed frequency band; and processing to output each image after performing the inverse Fourier transform.

[0022] (10) Further, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following process, the process comprising: processing to acquire multiple frame images; processing to calculate periodic changes characterizing the organ while maintaining the absolute positional relationship of each pixel in the multiple frame images; processing to extract pixels that vary with the frequency corresponding to the periodic changes characterizing the organ state in the image through a digital filter; and processing to output an image including the pixels extracted by the digital filter.

[0023] (11) Further, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following process, the process including: processing to acquire multiple frame images; processing to calculate periodic changes characterizing the organ while maintaining the absolute positional relationship of each pixel in the multiple frame images; processing to select a color corresponding to the periodic change rate characterizing the organ state; and processing to add the selected color with respect to the change rate of pixel values ​​and display the image on a display.

[0024] (12) Further, the diagnostic support program according to one aspect of the invention has the following features: classifying multiple frame images into multiple groups and calculating the periodic changes in the state of an organ that characterizes the state in which the relationship of each pixel in the multiple frame images belonging to each group is maintained.

[0025] (13) Further, the diagnostic support program according to one aspect of the invention has the feature of calculating image changes in block regions in which the indicated organs maintain the pixel-relative positional relationship between each frame of images, regardless of whether they are adjacent frames.

[0026] (14) Further, the diagnostic support procedure according to one aspect of the invention also includes a process for restoring the transmittance of a specific region in a frame image taken from X-rays to its original state after a change in transmittance.

[0027] (15) Further, the diagnostic support procedure according to one aspect of the invention also includes processing the signal values ​​of regions of magnetic field inhomogeneity in the frame images taken by MRI and converting them into images obtained when the magnetic field is homogeneous.

[0028] (16) Further, the diagnostic support program according to one aspect of the invention has the following features: regardless of whether they are adjacent frames, it calculates the rate of change from all aspects of the changes in the whole organ between each frame image, and calculates the changes in the image within a specific block region based on the calculated rate of change.

[0029] (17) Further, the diagnostic support procedure according to one aspect of the invention has the feature that the rate of change varies depending on the location of the block region in the organ or remains constant throughout the organ.

[0030] (18) Further, the diagnostic support program according to one aspect of the present invention has the following features: a process of setting a maximum outer edge when the organ size is at its maximum; a process of setting a minimum outer edge when the organ size is at its minimum; a process of calculating the outer edge coefficients of organs of other sizes in each image using the maximum outer edge and the minimum outer edge; and a process of displaying the waveform and waveform control points corresponding to the outer edge coefficients of organs in each image on a graphic, wherein the coefficients of each image are changed by changing the position of the control points.

[0031] (19) Further, the diagnostic support program according to one aspect of the invention has the feature of displaying the pixel average of an organ image graphically.

[0032] (20) Further, the diagnostic support program according to one aspect of the invention has the feature of displaying images of organs alongside the graphic.

[0033] (21) Further, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following processes, the processes including: processing of acquiring multiple frame images; processing of segmenting the human organ image in each frame image into multiple block regions; processing of calculating the image changes of the block regions in each frame image; processing of performing Fourier transform on the change values ​​of each block region; and processing of classifying each region by color based on the composition ratio of frequency components.

[0034] (22) Further, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following process, the process comprising: processing of acquiring multiple frame images; processing of extracting pixels with coordinates different from those of each pixel in a specific frame image and calculating periodic changes characterizing organ state for each pixel in a next frame image and subsequent frame images; processing of performing a Fourier transform on the periodic changes characterizing organ state; processing of extracting the spectrum from a fixed frequency band containing the spectrum corresponding to the frequency of organ movement to obtain the spectrum obtained after Fourier transform; processing of performing an inverse Fourier transform on the spectrum extracted from the fixed frequency band; and processing of outputting each image after performing the inverse Fourier transform.

[0035] (23) Further, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following process, the process comprising: processing of acquiring multiple frame images; processing of extracting pixels with coordinates different from those of each pixel in the next frame image and subsequent frame images for each pixel in a specific frame image and calculating periodic changes characterizing organ state; processing of extracting pixels that vary with the frequency corresponding to the periodic changes characterizing organ state in the image through a digital filter; and processing of outputting an image comprising the pixels extracted by the digital filter.

[0036] (24) Furthermore, according to one aspect of the present invention, a diagnostic support program is a diagnostic support program that analyzes images of human organs and displays the analysis results, the program causing a computer to perform the following process, the process comprising: processing for acquiring multiple frame images; processing for extracting pixels with coordinates different from those of each pixel in the next frame image and subsequent frame images for each pixel in a specific frame image and calculating a periodic change characterizing the organ state; processing for selecting a color corresponding to the periodic change rate characterizing the organ state; and processing for adding the selected color to the change rate of pixel values ​​and displaying the image on a display.

[0037] Effects of the present invention

[0038] According to one aspect of the invention, based on the waveform or frequency variation trend of the heart, or in its image, the human heart can be controlled as an object and images representing actual motion can be displayed. Attached Figure Description

[0039] Figure 1 This is a diagram showing the overall configuration of the diagnostic support system according to this embodiment.

[0040] Figure 2A It is a diagram showing a cross-section of the heart.

[0041] Figure 2B It is a diagram showing a cross-section of the heart.

[0042] Figure 2C It is a diagram showing a cross-section of the heart.

[0043] Figure 2D This is a diagram showing an example of the von Lund-Noël partitioning method.

[0044] Figure 3A It is a graph representing the "intensity" changes in a specific block and the results obtained through Fourier analysis.

[0045] Figure 3B It is a graph showing the Fourier transform results obtained by extracting the frequency components close to the heartbeat, and the changes in the "intensity" of the frequency components close to the heartbeat obtained by performing an inverse Fourier transform on them.

[0046] Figure 3C This is a diagram showing an example of extracting the spectrum obtained after Fourier transform from a specific fixed frequency band.

[0047] Figure 4 This is a flowchart illustrating the outline of the image processing in this embodiment.

[0048] Figure 5 This is a flowchart illustrating the outline of the image processing in this embodiment.

[0049] Figure 6 This is a flowchart illustrating the outline of the image processing in this embodiment.

[0050] Figure 7A This is a diagram showing the left lung of the human body from the front.

[0051] Figure 7B This is a diagram showing the left lung of the human body from the left side.

[0052] Figure 8A This is a diagram showing the left lung of the human body from the front.

[0053] Figure 8B This is a diagram showing the left lung of the human body from the left side.

[0054] Figure 9A This is a diagram showing the left lung of the human body from the front.

[0055] Figure 9B This is a diagram showing the left lung of the human body from the left side.

[0056] Figure 10A This is a diagram showing the left lung of the human body from the front.

[0057] Figure 10B This is a diagram showing the left lung of the human body from the left side.

[0058] Figure 11 This is a diagram illustrating an example of the lung region detection method of the present invention. Detailed Implementation

[0059] The inventors of this invention, recognizing the lack of practical application of visualization techniques for organ (e.g., the myocardium of the heart) movement, discovered that physicians could be supported in their diagnoses by representing not only deviations in organ movement but also their inactive positions, and thus proposed this invention. Specifically, because filtering was not well-applied in past cardiac image processing techniques, the inventors applied a filter to extract target frequencies from cardiac images, making previously difficult-to-observe parts visible. In this way, diagnoses that previously required significant skill can be simplified, and motion images can be qualitatively represented, thereby providing objectivity in the displayed content. Although the heart is used as an example organ in this specification, it is self-evident that the invention is not limited to the heart but can be applied to various organs and blood vessels.

[0060] In the invention of this application, it is assumed that myocardial motion is constant and can be distinguished by comparison with previous images of the subject, by comparison with the range to be analyzed and the average of the whole myocardium, by comparison with normal images, or by comparison with age samples.

[0061] First, the basic concepts of the invention are explained. According to the invention, regarding periodic changes characterizing the state of human organs, such as the cross-sectional area, surface area, and volume of the heart, relative to motion captured in a repetitive manner within a fixed cycle, a fixed repetitive or fixed motion (routine) is captured as a wave over the entire or a specific portion of the time axis and measured. For the measurement results of the wave, (A) the shape of the wave itself or (B) the wave interval (frequency: Hz) is used.

[0062] For images of the heart, there may be waves linked in a similar manner within the same time period. For example, in the case of a heartbeat, the following approximation can be conceptualized.

[0063] (Average density change over a rough range) ≈ (Heartbeat) ≈ (Changes in the heart) ≈ (Electrocardiogram) ≈ (Changes in heart surface area and volume)

[0064] In this invention, for example, for the heart, analysis can be performed by focusing on the deformation of the local ventricular wall (wall thickness expansion, wall thickness contraction), and systolic thickness, diastolic thickness, and wall motion (intima and adventitia) can be captured and periodically presented. Furthermore, for the heart, the "relative wall thickness (RWT)" calculated based on "(2 × posterior wall thickness) / left ventricular end-diastolic diameter)" and the "ejection fraction" calculated based on "(end-diastolic volume - end-systolic volume) / (end-diastolic volume)" can be captured and periodically presented.

[0065] In this invention, by using any of these data or by combining these data, images can be extracted with higher accuracy. In this case, multiple calculations may sometimes be performed interactively. At this time, artifacts related to the result are eliminated again, and function extraction is performed by extracting waveforms from new data, with the data waveforms becoming the first base, another modal waveform, etc., peripheral and multiple waveforms. In this case, the number of times can be one or more.

[0066] Here, when generating the basic data, the mutual component extraction is composed of multiple modes (e.g., a certain density, changes constituted by volume analysis, cardiac motion, etc.) or through multiple heartbeat waveform measurements, thereby improving accuracy. By doing so, artifacts can be reduced and accuracy improved based on some fixed prediction such as a straight line.

[0067] In this article, "density" is translated as "density," but in an image, it refers to the "absorption value" of pixels in a specific area. For example, in the case of CT scans, air, bone, and water are used as "-1000," "1000," and "0," respectively.

[0068] In this specification, "density" and "intensity" are used distinctly. As mentioned above, "density" refers to the absorption value and, in the original image of XP and XP motion images, exhibits high permeability. By digitizing the highly permeable portions as white, air, water, and bone are displayed as "-1000", "0", and "1000", respectively. On the other hand, "intensity" is a relative change from "density," for example, presented by normalizing the "conversion" to density and signal width. That is, "intensity" is a relative value of brightness, emphasis, etc. During the direct processing of the absorption value of XP images, it is represented as "density" or "density change (Δdensity)". Then, for the sake of image representation, it is converted as described above and represented as "intensity". For example, "intensity" is given in the case of color display of 256 levels of gray from 0 to 255. This terminology distinction applies to both XP and CT cases.

[0069] On the other hand, in the case of MRI, even if we try to set air, water, and bone to "-1000", "0", and "1000" respectively, the values ​​can vary significantly due to factors such as MRI pixel values, the type of measuring machine, the patient's physical condition during measurement, body structure, and measurement time. Furthermore, how MRI signals are acquired, such as T1-weighted images, also varies depending on the facilities and type of measuring machine, making a fixed approach impossible. Therefore, in the case of MRI, as with XP and CT, the definition of "density" cannot be applied. Thus, MRI processes relative values ​​from the initial extraction stage, representing them as "intensity" from the outset. Therefore, the processed signal is also "intensity".

[0070] Based on the above description, master data can be obtained. For the aforementioned master data, a new target to be measured is extracted within a fixed width and range of the waveform and Hz of the aforementioned master data. For example, only heartbeats are extracted, or extraction is performed within the width or range of frames representing the degree of vascular extraction. Furthermore, this waveform and Hz width are relatively and uniformly determined based on statistics obtained through multiple executions using waveform elements from other functions, human factors such as noise, waveforms of other modalities considered to have other adjustability and reproducibility, etc. Therefore, adjustments and experience (and machine learning can also be applied) are required. This is because as the width and range are expanded, elements of another function begin to enter, and if they are too narrow, elements of the function itself will be eliminated, thus requiring range adjustment. For example, in the case of multiple datasets, it is easy to specify the range, Hz, consistent width in the measurement, etc. Furthermore, fluctuations in the axis, width, range, and Hz caused by the extraction and width of mutual components can be estimated. That is, the axis setting of Hz is averaged through multiple superpositions, and the optimal range of each of the axis, width, range, and Hz is calculated using variance. At this point, there is a situation where the Hz (noise) of other behaviors is extracted, and if its wave exists, the relative measurement is also of the degree to which the wave is not contained in the presence of the wave.

[0071] Next, for new target data to be measured, the consistency rate and other inconsistencies with the captured waveform and Hz are quantified to calculate the numerical value for auxiliary diagnosis. For example, it can be applied to diagnostic aids by measuring the waveform consistency rate of the owner's disease and calculating the disease waveform consistency rate together with noise elimination from a pulse oximeter or auscultation. In this specification, the use of an adjustable consistency rate is described.

[0072] [Regarding Adjustable Consistency Rate]

[0073] In this specification, the trend of image variation is interpreted as an adjustable consistency rate. For example, a myocardial region is detected and segmented into multiple block regions to calculate the average density (pixel value x) of the block regions in each frame image. Then, the ratio (x') of the average pixel value of the block regions in each frame image to the width of the change (0%-100%) of the average density (pixel value x) from the minimum to the maximum value of the block regions is calculated. On the other hand, by using the ratio (y') of the change (y) in the myocardium of each frame image to the width of the change (0%-100%) from the minimum position to the maximum position of the myocardium, only block regions whose ratio (x' / y') falls within a predetermined fixed range are extracted.

[0074] Here, the cases where y′=x′ or y=ax (where a represents the numerical value of myocardial amplitude or the coefficient of density value) indicate complete consistency. However, this does not mean that only cases of complete consistency represent meaningful values; rather, values ​​with a certain fixed width should be extracted. Therefore, according to one aspect of this embodiment, the fixed width is determined using logarithms (log) as described below. That is, when calculated as a percentage (%) of the case where y=x, complete consistency of adjustability is “logY′ / x′=0”. Furthermore, for example, when the extracted range of adjustable consistency rate is narrow or (numerically narrow), it is determined to be “logY' / x'=-0.05~+0.05” in the range close to 0, and when the extracted range of adjustable consistency rate is wide or (numerically wide), for example, it is determined to be “logY' / x'=-0.5~+0.5” in the range close to 0. Since this range is narrow and the consistent values ​​are also high within this range, the adjustability can be considered to be higher. When this number is calculated by determining the ratio value for each pixel, a normal distribution with a peak value of perfect uniformity is obtained in the case of healthy individuals. Conversely, in those suffering from disease, the distribution of this ratio value is lost. Furthermore, as mentioned above, the method of determining the width using logarithms is merely an example, and the invention is not limited thereto.

[0075] In other words, this invention performs "image extraction" as follows: (average density change within a rough range) ≈ (heartbeat) ≈ (cardiac changes) ≈ (electrocardiogram) ≈ (changes in heart surface area and volume), and is also applicable to methods other than logarithmic methods. This method allows for the display of frequency-adjustable images.

[0076] In the case of blood vessels, the series of density changes (x (a waveform in the hilum region), with a slight time delay (phase change)) resulting from a series of contractions of the heart (y) are present as is, thus appearing as y = a'(Xt). In the case of perfect consistency, since t = 0, then y = x or y = a'x. Here, when the range of adjustable consistency extracted is narrow or (numerically narrow), for example, defined as "log Y' / x' = -0.05 to +0.05" in a range close to 0, and when the range of adjustable consistency extracted is wide or (numerically wide), for example, defined as "log Y' / x' = -0.5 to +0.5" in a range close to 0, the adjustable rate can be said to be higher because this range is narrower and the consistent value is also higher in this range.

[0077] In the case of other blood vessels, the aforementioned "part responding to the heart" is excluded, and the density drawn from the central side of the hilum can be used. The case of peripheral blood vessels can be treated similarly.

[0078] Furthermore, the present invention can also be applied to the circulatory system. For example, changes in cardiac density are directly correlated with changes in blood flow density towards the hilar region (peripheral lung field), and a series of changes in cardiac density are transformed into changes in hilar density, which are then transmitted to the lung field as is. This appears to be achieved by obtaining a small phase difference from the relationship between changes in cardiac density and hilar density. Furthermore, since density changes in the hilar region, etc., are correlated as is with changes in lung field blood flow density, adjustability can be represented by a constant rate of 1 (consistency relationship when Y≈X) reflecting the constant rate. Additionally, for the cervical vascular system, density changes plotted at the central cardiac vessels surrounding it appear to be directly correlated, or similarly correlated with a slight phase. Then, as density changes according to the background and propagates, this can be considered a case where an adjustable consistency rate propagates density changes.

[0079] Here, for each of the amount of change in an image and the rate of change in an image, when the amount of change in heart density is set to 1, in order to be displayed as a relative value (standard difference signal density / intensity), the amount of change and the rate of change can be extracted for each of the following cases: (1) for each different image of each image, when each image is set to 1 (generally assumed), (2) for each different image of each image, the ratio of heart rate set to 1 obtained by adding density (amount of change and rate of change) to it, and (3) for taking multiple pictures, the ratio obtained as the total density at each breath is set to 1.

[0080] Furthermore, in the case of 3D imaging such as MR, the difference in intensity or density of the heartbeat, obtained by summing the intensity (in the case of MR) or density (in the case of CT) (when it is set to 1), can be converted into "peak flow data" of the heartbeat (at rest, or even under load). For this value, the actual measured cardiac workload and rate of motion can be converted into this value by calculating the ratio of intensity or density when using at least MRI, CT, etc., to calculate "3D × time". Similarly, it is possible that the distribution of the "capillary phase" of "flow" in the lung field presents an estimate of the distribution of peripheral pulmonary blood flow or volume converted from a single cardiac output.

[0081] That is, it satisfies (average of density changes within a rough range) ≈ (one heartbeat) ≈ (one cardiac change) ≈ (one electrocardiogram) ≈ (one change in heart surface area and volume), and when only 10% or 20% of the change is taken out, the estimate can be calculated by (the total of these) × (the change at that time).

[0082] Then, for new target data to be measured, diagnostic images are calculated by imaging the already captured waveforms and the consistency rate or other inconsistencies in Hz. For example, the difference between normal swallowing and the patient's swallowing can be visualized, and the difference between the movements the patient has completed so far and the movements the patient is currently performing can be shown. For example, variations and differences in walking and swinging patterns can be included.

[0083] The extracted changes are visualized and plotted on the image. This is cardiac function analysis and vascular (blood flow) analysis, as described below. Then, the rate of change in the myocardium is visualized. At this point, in some cases, artifacts related to the results are eliminated again, and function extraction is performed by extracting the data waveform as the first basis, other modal waveforms, peripheral and multiple waveforms from the new data waveform. The methods for artifact elimination will be described later.

[0084] Furthermore, there is another situation where characteristic quantities can be grasped even from those extracted from the variable components other than those extracted as described above. For example, when grasping the movement of the abdominal intestine, attempts are made to extract the movement of the abdominal intestine by excluding the effects from abdominal respiration and blood vessels.

[0085] Furthermore, based on the rate of change caused by extraction, images requiring specific fixed capture times (CT, MRI, special X-ray photography, PET / scintillation scans, etc.) are corrected to provide clearer, more accurate images. For example, it can be used to improve aortic heart correction, cardiac morphology correction, bronchial blurring correction, perithoracic evaluation, and imaging when patients cannot hold their breath (which may require several minutes of capture time).

[0086] Hereinafter, embodiments of the present invention will be explained with reference to the accompanying drawings. Figure 1 This diagram shows a schematic configuration of the diagnostic support system according to this embodiment. The diagnostic support system performs a specific function by having a computer execute a diagnostic support program. Basic module 1 includes a cardiac function analysis unit 3, a blood flow analysis unit 5, another blood flow analysis unit 7, a Fourier analysis unit 9, a waveform analysis unit 10, and a visualization / digitalization unit 11. Basic module 1 acquires image data from a database 15 via an input interface 13. The database 15 stores images, for example, via DICOM (Digital Imaging and Communication in Medicine). The image signal output from basic module 1 is displayed on a display 19 via an output interface 17. Next, the function of the basic module according to this embodiment will be explained. Furthermore, the input image is not limited to the database 15, but can also be input via the input interface 13 or actively from other external devices.

[0087] [Refined implementation of dynamic region detection]

[0088] Contrast in dynamic regions such as the lung field, thoracic cavity, and heart is sometimes non-uniform along a straight line. In such cases, the shape of the dynamic region can be detected more accurately by changing the threshold used for noise cancellation and performing the detection process multiple times. For example, for the left lung, the contrast along the diaphragm line tends to be weaker closer to the interior of the body. There are also cases where the apex and base of the heart have less motion, while the central part of the heart has greater motion. In these cases, the remaining part of the left half of the diaphragm, the parts of the heart with greater motion, and the parts of the heart with less motion can also be detected by changing the setting of the threshold used for noise cancellation or by multiplying the pixel values ​​by different factors. By repeating this process multiple times, the shape of the entire diaphragm or the entire heart can be detected. In this way, this method can also digitize not only the position of the diaphragm, but also the rate and amount of change of lines and surfaces related to the shape of the chest, heart, and dynamic regions, and it can be used for new diagnoses.

[0089] In this way, the location or shape of the diaphragm and heart detected during diagnosis can be utilized. That is, the graphical coordinates of the diaphragm and heart can be plotted; the coordinates of the thoracic cavity, diaphragm, and heart can be calculated using curves (surfaces) or straight lines calculated as described above; and heartbeats, vascular pulsations, lung field "density," etc., can be plotted as locations corresponding to periodicity or coordinates. This method can be applied to dynamic regions related to breathing and heartbeats.

[0090] Such a method allows for measurements in response to changes in frequency bands, not only in the Hz of exhaled air, inhaled air, systole, and diastole, but also in the frequency (Hz) of the dynamic region associated with the diaphragm or respiration, or the frequency (Hz) of the cardiac dynamic region. Then, during the spectral extraction of the BPF (bandpass filter), the BBF can be set within a fixed range according to the various states of respiration or the heart; axial changes in the BPF position during each "reconstruction phase" of respiration or the heart can lead to an optimal state; and the BPF is prepared to accompany these changes. Even if the respiratory rhythm changes, such as slow breathing or cessation of breathing (Hz=0), or even if transient cardiac fibrillation (very high frequency) or cessation of breathing occurs (Hz=0), images can be provided based on the foregoing.

[0091] Additionally, the frequency of the entire exhaled or inhaled air can be calculated based on the ratio of respiratory elements (including respiratory elements of all or part of exhaled or inhaled air) to the total exhaled or inhaled air. Similarly, systolic or diastolic, frequency elements, and other overall frequencies can be calculated based on cardiac dynamic elements (including cardiac dynamic elements of all or part of systole or diastole) to the ratio of cardiac systole and diastole, a single heartbeat, and the entire measured heartbeat. Furthermore, multiple diaphragm and cardiac measurements can be performed, and those with stable signals and waveforms can be selected. Therefore, at least one frequency of respiratory and cardiac elements can be calculated based on the detected position or shape of the diaphragm or the position or shape of the dynamic region associated with breathing, and the frequency representing the heartbeat can be calculated. When the position or shape or dynamic region of the diaphragm and heart is to be / can be determined, the frequency of respiratory and cardiac elements and the heartbeat can be determined. Even if a portion of the waveform is segmented, the method can track subsequent waveforms. Therefore, even if the frequency of a respiratory or cardiac element changes midway, the original respiratory or cardiac element can be followed. In addition, sudden changes often occur in heartbeats, but the same principle can be applied to the heart vessels and organs associated with heart vessel waveforms.

[0092] [Lung Field Detection]

[0093] In this invention, lung field detection refinement can be implemented as an aspect of the aforementioned "dynamic region detection refinement." During this process, after setting the maximum and minimum lung fields, these values ​​are used to calculate other lung fields. Figure 11 This diagram illustrates an example of the lung field detection method of the present invention. In this method, a "B-spline curve" is used to represent the "coefficients of each image." Figure 11 In the diagram, waveform X represents the coefficient of each image L displaying the lung field, and from left to right, they represent the coefficients of the first image, the second image, and so on. Figure 11As the control point Y moves, the coefficients of each image change smoothly. In this invention, the graph of the coefficients can be directly edited in this way. Figure 11 In this diagram, the "gray polyline Z" represents the "average pixel value for each image." When captured under optimal conditions, the change in lung field size corresponds to the change in the average pixel value. In this paper, this average pixel value is smoothed using curve fitting and used directly as a "coefficient." The same method can be applied to the heart and other organs involved in cardiovascular frequencies.

[0094] [Cardiac Function Analysis]

[0095] Figure 2 is a cross-sectional view showing the general structure of the heart. "Cardiac function" is generally defined as "the pumping function of the left ventricle, which circulates blood throughout the body." Cardiac function analysis is important for evaluating the prognosis of patients with ischemic heart disease, especially myocardial infarction. For example, if the left ventricular ejection fraction (EF) is low, the heart's pumping output is reduced, and it cannot pump enough blood throughout the body. Other cardiac functions include left ventricular end-diastolic volume (EDV), left ventricular end-systolic volume (ESV), stroke volume (SV), cardiac output (CO), and cardiac index (CI). For local evaluation of the myocardium, a "bull's eye map" is used, which displays wall thickness, wall motion, rate of change of wall thickness, etc. Figure 2B and 2C As shown, Figure 2B and 2C It is perpendicular to Figure 2A A cross-section of the plane along the central axis A. This "bull's-eye" image is an image in which the apical cross-section is placed at the center, while the short-axis tomographic images are arranged in concentric circles outside the center, with the basal cross-section at the outermost position.

[0096] According to this embodiment, in addition to using a "bull's-eye diagram," the periodicity of cardiac motion is analyzed based on the following indicators: The periodicity of cardiac motion is analyzed by using density / intensity in a fixed region within the cardiac region. Furthermore, data obtained by another measurement method, such as spirometry, can be used, comprising a range of fixed volumetric densities / intensities measured in areas exhibiting high X-ray permeability (in addition to various modalities including CT and MRI), as well as external input information. Additionally, it is desirable to compare the analysis results for each heartbeat and analyze trends from multiple data points to improve data accuracy. Furthermore, the cardiac border can be identified, and frequency can be obtained based on changes in the cardiac border. Furthermore, the boundaries of the lung fields can be identified, and frequency can be obtained from the migration of these boundaries.

[0097] [Vascular Pulsation Analysis]

[0098] According to this embodiment, vascular pulsation is analyzed based on the following indicators: Vascular pulsation is analyzed using density / intensity changes in each region, either from measurements of other modalities such as electrocardiograms and pulse oximeters, or by specifying the location of the heart / hilum / aorta from the lung contour. Furthermore, density / intensity changes in the target region are analyzed by manually performing plotting on the image. Therefore, heartbeat elements obtained from heartbeats or vascular pulsation can also be used. Furthermore, improving data accuracy is desirable by comparing the analysis results of each pulse and analyzing the trends of multiple data fragments. Moreover, accuracy can be improved by performing density / intensity extraction of each region multiple times, and by performing the above operations for a fixed range. Additionally, there is a method for inputting cardiovascular pulsation frequency or band.

[0099] [Identification of the Heart Region]

[0100] Images are extracted from the database (DICOM), and cardiac regions (especially myocardium) are automatically detected using the results of the cardiac function analysis described above. The myocardium is then segmented into multiple block regions to calculate the variations in each region. Here, the size of the block region is determined based on the shooting speed. When the shooting speed is slow, it is difficult to specify the corresponding region on subsequent frames, resulting in a larger block area. On the other hand, with a fast shooting speed, the number of frames per unit time is high, so tracking can be performed even with a small block area. Furthermore, the size of the block region can be calculated depending on the timing of the selected cardiac motion cycle. Here, it is often necessary to correct for deviations in the myocardial region. In this case, cardiac motion is identified, and the relative position of the cardiac contour is further determined, with a relative evaluation made based on this motion. Additionally, when the block region is too small, flickering often occurs in the image. To prevent this, the block region needs to have a fixed size.

[0101] [Preparation of Block Regions]

[0102] Next, the method of dividing the myocardium into multiple block regions will be explained. Figure 2B and 2C This diagram illustrates a method for radially segmenting the myocardium from the center of the heart. For the cardiac region, the positional relationship between cardiac motion and blood vessels is determined, and the relative position of the cardiac contour is determined, with an evaluation based on this relative motion. Therefore, in this invention, after automatically detecting the cardiac contour, the myocardial region is segmented into multiple block regions, and the image variation values ​​(pixel values) included in each block region are averaged. Thus, even if the heart morphology changes over time, changes in the region of interest can be tracked over time.

[0103] On the other hand, when segmenting into block regions without specifying the heart region, the region of interest falls outside the heart region due to changes in the heart over time, resulting in meaningless images. Furthermore, there are methods that input heartbeats or frequency bands. These methods are also applicable to 3D stereo images. Region segmentation in 3D stereo images can be calculated by keeping the pixels constant. This relative evaluation based on relative positional migration can be performed between adjacent frames or for every integer multiple of images, such as every two or three images. Furthermore, multiple images can be grouped together and processed for each group.

[0104] Figure 7A It is a frontal view of the human left lung, and Figure 7B This is a diagram showing the left lung of the human body from the left side. Figure 7A and Figure 7B Both indicate inhalation, that is, the lungs during inhalation. Figure 8A It is a frontal view of the human left lung, and Figure 8B This is a diagram showing the left lung of the human body from the left side. Figure 8A and 8B All figures show exhalation, i.e., the lungs during exhalation. As these figures show, the morphology of the lung fields changes significantly during respiration, but the rate of change is greater on the side of the diaphragm and smaller on the opposite side. In this invention, the position of each region within the lung field varies according to this rate of change. This allows for relative evaluation based on the relative positional relationships of each region within the lung field. Furthermore, based on the rate of change of the lung fields (e.g., the average rate of change), the relative positional relationships can be represented by a constant rate of change of the lung field regions, or the rate of change can be adaptively varied within the lung field regions according to their distance from the diaphragm. In this way, by using the rate of change of the lung field regions, an image synchronized with the respiratory cycle can be displayed.

[0105] Figure 9A It is a frontal view of the human left lung, and Figure 9B This is a diagram showing the left lung of the human body from the left side. Figure 9A and 9B All of them show inhalation, that is, the lungs in an inhalation state. Figure 10A It is a frontal view of the human left lung, and Figure 10B This is a diagram showing the left lung of the human body from the left side. Figure 10A and 10B Both indicate exhalation, that is, the lungs in the exhalation state. For example, as Figure 9A and 9B As shown, marker P1 is drawn at the location in the lung field region during inspiration. It is assumed that marker P1 is a fixed point defined by two-dimensional coordinates, and that these coordinates remain unchanged even during expiration; therefore, marker P1 exists at the same location, as shown. Figure 10Aand Figure 10B As shown. On the other hand, in this invention, as described above, the evaluation is based on the relative positional relationship with the entire lung field region; therefore, during exhalation, the point migrates to the position marked P2 instead of the position marked P1. Alternatively, vector evaluation can be used at this time to determine the point to be moved to by comparing the point drawn during inspiration with the point determined during exhalation.

[0106] To divide a region, Thiessen partitioning can be used. For example, Thiessen partitioning... Figure 2D The diagram illustrates a method for segmenting the nearest neighbor region of each base point by drawing perpendicular bisectors on straight lines connecting adjacent base points. Applying this von Ronoy segmentation reduces computation time. Furthermore, when drawing straight lines connecting adjacent base points, weights can be applied based on the analysis objective. For example, when segmenting the pulmonary artery region, heavier regions can be weighted more heavily, while lighter regions can be weighted less heavily. This allows segmentation to be tailored to the analysis objective while reducing processing overhead. Additionally, multiple block regions resulting from segmentation can be classified based on indicators such as pixel value variations (periodic changes).

[0107] In this approach, after dividing a region into multiple block regions, the image changes in each block region are calculated based on its relative position to a dynamic region such as the heart. In this paper, not only is the block extent itself considered a unit for obtaining signal differences, but these differences can also be taken from a range smaller than the block or a larger range surrounding the block. Furthermore, the vertical extent can be added only near the diaphragm, the horizontal extent only in other dynamic regions, or the shape of the extent can be distorted, or pixel regions can be connected. Moreover, after calculating one or more differences, it is desirable to redefine the block shape to match the shape of the entire lung field and heart. For example, after processing from the first image to the second image, the shape of the block can be recreated in the second image to match that shape, and then the second image can be compared with the third image.

[0108] The above description describes the "relative positional relationship" considering organ movement, but the invention is not limited to this and can also perform image processing while maintaining the "absolute positional relationship of each pixel" across multiple frames. The "absolute positional relationship of each pixel" refers to the relationship between pixels whose coordinates are specified based on a two-dimensional coordinate axis defined on a frame image. In other words, it is a pixel processing method where the pixels of interest remain constant. The processing that maintains this "absolute positional relationship of each pixel" assumes multiple frame images, but the number of frame images is not specified. Multiple frame images can be divided into multiple groups, and each group can include the same number of frame images, or each group can include a different number of frame images.

[0109] In other words, by dividing multiple frame images into multiple groups and maintaining the absolute positional relationship of each pixel in the multiple frame images belonging to each group, the changes in the pixel value of the organ, the changes in the distance from the center of the organ to the outer edge, or the changes in the volume of the organ are calculated. This allows even slight changes in pixel values ​​to be treated as equal values, thereby reducing the amount of data and processing steps.

[0110] Furthermore, a positional relationship that is neither relative nor absolute can be envisioned. This means that a point P with specific coordinates can be defined in a particular frame image, and another point Q with different coordinates can be defined in the next and subsequent frames, but in this case, the magnitude of the vector PQ assumes that the motion relative to the organ is small. Additionally, for point Q, other points R with slightly different coordinates are defined in the next and subsequent frames. By repeating this operation, other points slightly deviating from the specific point P are extracted for each frame image, and this implementation method is applied. Specifically, multiple frame images are acquired, and for each pixel in a particular frame image, pixels with different coordinates than each pixel in the next and subsequent frames are extracted, and the periodic changes characterizing the organ state are calculated.

[0111] Then, a Fourier transform is performed on the periodic changes representing the organ's state. The spectrum within a fixed frequency band, including the spectrum corresponding to the frequency of organ movement, is extracted to obtain the Fourier transform spectrum. An inverse Fourier transform is then performed on the spectrum extracted from the fixed frequency band, and each image after the inverse Fourier transform is output. Furthermore, pixels that change in accordance with the frequency representing the periodic changes in the organ's state in the image are extracted using a digital filter, and an image including the pixels extracted by the digital filter can be output. Additionally, a color corresponding to the rate of change representing the organ's state can be selected, and the selected color can be added to the rate of change of pixel values. The image can then be displayed on a monitor. This allows for the representation of organ movement.

[0112] Next, artifacts are eliminated to insert the image data. That is, when bones and the like are included in the analysis, they appear as noise, so it is desirable to remove this noise by using a noise cutoff filter. For X-ray images, air and bones are traditionally set to -1000 and 1000 respectively, so highly penetrating areas have low pixel values ​​and are displayed as black, and low-penetrating areas have high pixel values ​​and are displayed as white. For example, when displaying pixel values ​​in 256 grayscale, black becomes 0, and white becomes 255.

[0113] Within the cardiac region, X-rays have difficulty penetrating areas containing blood vessels and bones, resulting in high pixel values ​​and a washed-out appearance in X-ray images. The same applies to other CT and MRI scans. In this paper, based on the results obtained through the aforementioned cardiac function analysis, artifacts can be eliminated by interpolating data using values ​​within the same phase, based on the waveform of each heartbeat. Furthermore, when "coordinates differ from the original," "pixel values ​​change significantly," or "frequency and concentration are abnormally high," these are stopped out, and the remaining acquired images can be easily used for Hz calculations of the heart and adjustments to the myocardial region, for example, by identifying continuous, smooth waveforms using methods such as least squares. Additionally, when overlaying images, there are two methods: (1) overlapping the acquired comparative image obtained from each image before and after acquisition with the coordinates themselves, and (2) overlaying the positional information onto a reference by extending the image to a reference after each image before and after acquisition. By using the above methods, the shape of the cardiac region can be corrected, and image variations in the block region can be corrected.

[0114] Here, we explain "reconstruction" on the timeline. For example, when the inhalation time is 2 seconds at 15 f / s, 30+1 images are obtained. In this case, if we simply overlay 3 images each time, we can "reconstruct" for every 10%. For example, when 0.1 seconds represents 10%, and the images are only captured at 0.07 seconds and 0.12 seconds, a 0.1-second "reconstruction" is needed. In this case, the intermediate value before or after 10% in the image, a value (the average of the two), is assigned for "reconstruction." Furthermore, the timeline can be taken out, and this coefficient can be changed according to the time ratio. For example, when there is no 0.1-second capture value due to the timeline difference, but capture times of 0.07 seconds and 0.12 seconds, the "recalculation" is "(its 0.07-second value) × 2 / 5 + (its 0.12-second value) × 3 / 5" for "reconstruction." Furthermore, it is desirable to include a "maximum differential intensity projection" of 0-100%, and to calculate this by providing ranges such as 10%-20% "reconstruction," 10%-40% "reconstruction," etc. In this way, the un-captured portions can be "reconstructed" at a rate equivalent to a heartbeat. Additionally, according to the invention, "reconstruction" can also be performed on a series of movements similar to the heart, blood flow, and those associated with these.

[0115] [Fourier Analysis]

[0116] Based on the periodicity of cardiac motion and vascular pulsation as described above, Fourier analysis was performed on the density / intensity values ​​and their changes in each block region. Figure 3A It is a graph representing the intensity changes in a specific block and the results of its Fourier analysis. Figure 3BThis is a graph representing the Fourier transform results of the intensity changes of the frequency components near the heartbeat, obtained by extracting them and performing an inverse Fourier transform. For example, when performing a Fourier transform (Fourier analysis) on the intensity changes in a specific block, the result is as follows... Figure 3A As shown in the figure. Then, these results are as follows: Figure 3B As shown on the right, by means of Figure 3A This is obtained by extracting frequency components close to the heartbeat from the frequency components shown. By performing an inverse Fourier transform on this, the intensity variation that matches the heartbeat can be obtained, such as... Figure 3B As shown on the left.

[0117] Here, when performing an inverse Fourier transform on a spectrum including frequency components, the inverse Fourier transform is performed by simultaneously considering the frequency elements (heartbeat and cardiovascular pulsation frequencies) and frequency bands (which may be BPF) specified by the density of heartbeat and blood flow; or based on either element.

[0118] Furthermore, the AR method (Autoregressive Moving Average model) can be used to complete calculations in a short time when performing Fourier transforms. According to the AR method, the autoregressive moving average model employs the Yule-Walter equation or Kalman filter, and this computation can be compensated for by using derived Yule-Walter, PARCOR, or least squares estimation methods. This allows for near real-time image acquisition to aid calculations and correct artifacts at a higher speed. Through this Fourier analysis, the properties of the image can be extracted and displayed in each block region.

[0119] Here, by performing a Fourier transform on the changes in each block region of each frame image, the spectrum in a fixed frequency band, including the spectrum corresponding to the cardiac cycle, can be extracted to obtain the spectrum obtained after the Fourier transform. Figure 3C This is a diagram illustrating an instance of extracting a fixed frequency band from outside the Fourier transform spectrum. For the frequency f of the composite spectrum, the frequencies f1 (heartbeat component) and f2 (pathological blood flow component) of each composite source satisfy the relationship "1 / f = 1 / f1 + 1 / f2", and the following methods can be used when extracting the spectrum.

[0120] (1) Extract heartbeats with high spectral ratio.

[0121] (2) The spectrum is extracted by segmentation between the peak of the spectrum corresponding to the heartbeat / pathological blood flow and the peak of multiple adjacent composite waves.

[0122] (3) The spectrum is extracted by segmentation at the peak and valley portions of the spectrum corresponding to the spectrum of heartbeat / pathological blood flow and the spectrum corresponding to multiple adjacent composite waves.

[0123] (4) Extract the spectrum contained in a fixed bandwidth from the heartbeat component (blood flow component). In this case, a spectrum with multiple overlapping spectra is obtained, but each spectrum can be recovered by separating each component.

[0124] As described above, according to the present invention, this does not refer to using a fixed BPF, but rather to extracting the spectrum from a fixed frequency band that includes the spectrum corresponding to the periodicity of heart movement. Furthermore, according to the invention of this application, frequencies other than the frequencies of heart movement obtained from the frame image (e.g., further, density / intensity in each region, and heartbeat elements obtained from heartbeats or vascular pulsations) can be extracted from the spectrum obtained after Fourier transform, or from the spectrum in a fixed frequency band that includes the spectrum corresponding to the frequency input by the operator from an external source (e.g., a spectral model).

[0125] Here, the composite spectral elements become 50%+50% in the case of only two components (heartbeat and pathological blood flow), and in the case of three components, the distribution becomes one-third for each. Thus, the composite spectrum can be calculated to some extent from the percentage of the heartbeat component spectrum and the percentage of the pathological blood flow component spectrum, along with their spectral components and height. The spectrum can be extracted based on its high ratio (%). That is, the ratio of pathological blood flow component / heartbeat component to composite spectral component is calculated, and the spectral values ​​with high pathological blood flow component / heartbeat component are calculated and extracted. Furthermore, for diaphragm identification, in some cases, the spectrum corresponding only to regions with relatively constant Hz (frequency) (i.e., regions with small Hz variations) or their superimposed spectrum is extracted from data obtained by acquiring the frequencies of heartbeat and cardiac vessels. Moreover, in cases where spectral bands are determined, there are some cases where spectral bands and their surrounding regions within the range that produce Hz variations are identified. Therefore, not only can cases perfectly consistent with the periodicity of cardiac motion or vascular pulsation be extracted, but also the spectra that should be considered can be extracted, contributing to image diagnosis.

[0126] Furthermore, it is known that "heartbeat" and "respiration" fall within specific frequency bands. Therefore, by using a filter for respiration, for example, "0-0.5 Hz (respiratory rate 0-30 breaths / minute)," and for the circulatory system, for example, a filter for the circulatory system, "0.6-2.5 Hz (heartbeat / pulse rate 36-150 beats / minute)," the respiratory and circulatory frequencies can be pre-specified using these filters. This allows for the display of frequency-adjustable images. This is because sometimes changes in respiratory (lung) density are detected when acquiring changes in cardiac density, and changes in cardiac density are detected when acquiring changes in lung density.

[0127] [Waveform Analysis]

[0128] Waveform analysis is performed on cardiac, vascular, and electroencephalogram (EEG) data that are identified as constant waveforms during the examination. This includes repetitive movements in a constant state, such as foot movements. Furthermore, an analysis is performed to determine if there is a consistent trend by superimposing the Hz values ​​of the repetitive movements. The waveform data are compared, and the consistency rate between the two sets of data is calculated. Then, the data after Fourier analysis are compared.

[0129] [Digital Filter]

[0130] In addition, digital filters can be used instead of the Fourier analysis described above. Digital filters perform transformations between the time and frequency domains based on the Fast Fourier Transform and Inverse Fast Fourier Transform. These mathematical algorithms are used to extract the frequency components of the signal for adjustment. This achieves the same effect as the Fourier analysis described above.

[0131] [Visualization / Digitalization]

[0132] The results of the above analysis are visualized and digitized. As a standard uptake, the value is typically displayed relative / logarithmically by setting the average density / intensity of the entire lung field to 1. Furthermore, since only blood flow direction is considered, variations in that specific direction are usually removed. This allows only data from significant methods to be extracted. Pseudo-coloring is performed using the identification results of the cardiac region, varying with the analysis range. That is, the analysis results for each individual (subject) are fitted to a relative region according to a specific shape (minimum, maximum, mean, median) fitted to the phase. Furthermore, deformation of specific traits / phases allows for comparison of multiple analysis results.

[0133] Furthermore, when preparing a "standard heart," the relative positional relationships within the heart (myocardium) are calculated using the aforementioned analysis of cardiac motion. The "standard heart" is prepared by using lines that comprehensively average the cardiac contours, densities, etc., of multiple patients. Moreover, this concept is not limited to the heart but can be applied to the lungs (standard lungs) and other organs (standard organs). For example, "organ models" can be created based on age, sex, country, and disease severity.

[0134] In addition to the changes in heart pixel values ​​mentioned above, the distance from the heart center to the myocardium can also be calculated. Figure 2B and 2C Fourier analysis was performed on the changes in distance L shown, and in addition, Fourier analysis could be performed by calculating the changes in heart volume.

[0135] After preparing the "standard heart," as described above, adjustability, consistency rate, and inconsistency rate can be digitized and presented (display of a frequency-adjustable image). Furthermore, deviations from the normal state can be displayed. According to this embodiment, by performing Fourier analysis, the possibility of new diseases can be identified, along with self-comparisons to the normal state, and comparisons of the hands and feet with the other hand and foot on the opposite side. Furthermore, by digitizing adjustability, abnormalities in foot movement, swallowing, etc., can be identified. Moreover, it can be determined whether a person in a disease state changes after a certain period of time, and furthermore, if changes occur, the state before and after the change can be compared.

[0136] [Drawing of the Heart]

[0137] This specification employs a method to adjust the heart to achieve high matching accuracy by temporarily drawing the heart outline using a combination of Bezier curves and straight lines. For example, when the heart outline is represented by four Bezier curves and one straight line, the heart outline can be drawn by finding five points and four control points on the heart outline. Multiple heart outlines can be drawn by moving the position of a point, and the heart outline can be detected with high precision by using methods such as "making the total density value within the outline become the maximum value" and "making the difference between the density sum of several pixels inside and outside the outline become the maximum value". In addition, points near the outer edge can be extracted by classical binarized outline extraction, and the position of the control point of the Bezier curve can be adjusted by methods such as least squares. Furthermore, the above method is not limited to the heart and can be applied to other organs as "organ detection". This is applicable not only to planar images but also to stereoscopic images (3D images). By defining the surface equation and setting its control points, targets surrounded by multiple surfaces can be assumed to be organs.

[0138] [Cardiovascular Function Analysis Using Fourier Analysis]

[0139] Next, the cardiac function analysis using Fourier analysis according to this embodiment will be explained. Figure 4This is a flowchart showing an overview of the cardiac function analysis in this embodiment. Basic module 1 extracts images from the DICOM database 15 (step S1). Here, at least a plurality of frame images encompassing one heartbeat are acquired. Then, in each acquired frame image, at least in a fixed region of the myocardium, the cycle of cardiac motion is specified by using changes in pixel values, for example, changes in density (density / intensity) (step S2). Then, the cardiac (myocardial) region is detected (step S3), and the detected myocardium is segmented into multiple block regions (step S4). Here, as described above, the myocardium is radially segmented from the center of the heart using Thiessen segmentation. Then, the pixel value changes within each block region in each frame image are calculated (step S5). Hereinafter, the values ​​of the changes within each block region are averaged and represented as a single data point.

[0140] Furthermore, pixels can be blurred to some extent by dimming them to reveal the overall image. Specifically, in the case of blood vessels, low and high signal values ​​coexist, but if only the high signal values ​​can be roughly grasped, overall blurring is acceptable. For example, in the case of blood flow, only signals with a threshold or higher can be extracted. Specifically, using the values ​​in the table below as one pixel and obtaining the value at the center, by obtaining the proportion of the value at the center and averaging it within one pixel, its representation can be smoothly performed between adjacent pixels.

[0141] [Table 1]

[0142]

[0143] Furthermore, noise can be eliminated by cutoff for the varying values ​​within each block region. Next, based on the periodicity of cardiac motion described above, Fourier analysis is performed on the density / intensity values ​​and their variations in each block region (step S6). In this way, the properties of the image can be extracted and displayed in each block region.

[0144] Here, the spectrum from a fixed frequency band, including the spectrum corresponding to the periodicity of the heart, can be extracted to obtain the Fourier transform spectrum. Here, for the frequency f of the composite spectrum, the relationship "1 / f = 1 / f1 + 1 / f2" is satisfied between the frequencies f1 and f2 that constitute the composite source, and the following method can be used when extracting the spectrum.

[0145] (1) Extract cardiac motion with high spectral ratio.

[0146] (2) The spectrum is extracted by dividing the peak of the spectrum corresponding to heartbeat / blood flow into the middle of the peak of multiple adjacent composite waves.

[0147] (3) The spectrum is extracted by segmenting the peaks and troughs of the spectrum corresponding to heartbeat / blood flow and the spectrum of multiple adjacent composite waves.

[0148] Here, the composite spectral elements become 50%+50% in the case of only two components (heartbeat and pathological blood flow), and in the case of three components, the distribution becomes one-third for each. Thus, the composite spectrum can be calculated to some extent based on the percentages of the heartbeat component spectrum and the blood flow component spectrum, as well as their spectral components and heights. This spectrum can be extracted at its high ratio (%). That is, the ratio of blood flow component / heartbeat component to composite wave component is calculated, and the spectral values ​​with high blood flow component / heartbeat component are calculated and extracted.

[0149] Next, noise reduction is performed on the results obtained from the Fourier analysis (step S7). Here, truncation and artifact removal as described above can be performed. The above operations from step S5 to step S7 are performed at least once, and it is determined whether to complete (step S8). If not completed, proceed to step S5; and when completed, the results obtained from the Fourier analysis are displayed on the screen as a pseudo-color image (step S9). Alternatively, a black and white image can also be displayed. In some cases, repeating this process multiple times can improve the accuracy of the data. Therefore, the desired motion image can be displayed. Furthermore, the desired motion image can be obtained by correcting the image displayed on the screen.

[0150] In addition to the myocardial segmentation processing in steps S4 and S5 as described above, Fourier analysis can also be performed by calculating the change in distance from the center of the heart to the myocardium instead of steps S4 and S5.

[0151] [Cardiac Function Analysis Using Digital Filters]

[0152] Next, the cardiac function analysis using digital filters according to this embodiment will be explained. Figure 5 This is a flowchart showing an overview of the cardiac function analysis in this embodiment. Steps S1-S5 and S7-S9 are the same as those in the "Cardiac Function Analysis Using Fourier Analysis" described above, and are therefore omitted. Figure 5 In step T1, digital filter processing is performed (step T1). The digital filter performs a transformation between the time and frequency domains based on "Fast Fourier Transform and Inverse Fast Fourier Transform," and these mathematical algorithms are used to extract the frequency components of the signal for adjustment. This achieves the same effect as the Fourier analysis described above.

[0153] [Cardiovascular Function Analysis Using Adjustable Consistency Rate]

[0154] Next, the cardiac function analysis using adjustable consistency rate according to this embodiment will be explained. Figure 6This is a flowchart showing an overview of the cardiac function analysis in this embodiment. Steps S1-S5 and S7-S9 are the same as those in the "Cardiac Function Analysis Using Fourier Analysis" described above, and are therefore omitted. Figure 6 In step R1, an analysis of the adjustable consistency rate is performed (step R1). Therefore, the heart (myocardium) region is detected (step S3), and after segmenting the myocardium into multiple block regions (step S4), the average density (pixel value x) of the block regions in each frame image is calculated, and the ratio (x') of the average pixel value of the block regions in each frame image to the width of the change in average density (pixel value x) from minimum to maximum (0% to 100%) is calculated (step S5). On the other hand, the ratio (y') of the heart change (y) in each frame image to the width of the change in heart surface area (or volume) from minimum to maximum (0% to 100%) (x' / y') is calculated (step S5). By using these, only block regions whose ratio (x' / y') falls within a predetermined fixed range can be extracted (step R1).

[0155] Here, the case of y' = x' or y = ax (where a represents a numerical value of the surface area or volume of the heart, or a coefficient of the density value) indicates perfect consistency. However, this does not mean that only cases of perfect consistency represent meaningful values, and values ​​with a certain fixed width should be extracted. Therefore, according to one aspect of this embodiment, the fixed width is determined using logarithms (log) as described below. That is, when calculated as a percentage (%) of the case of y = x, perfect consistency of adjustability is "logY' / x' = 0". Furthermore, for example, when the range of the adjustable consistency rate is narrow or is a (numerically narrow) range, it is determined to be "logY' / x' = -0.05 to +0.05" within a range close to 0, and when the range of the adjustable consistency rate is wide or is a (numerically wide) range, for example, it is determined to be "logY' / x' = -0.5 to +0.5" within a range close to 0. Since this range is narrower, and the consistent values ​​within this range are also higher, the adjustability can also be said to be higher. When the quantity is calculated by determining the ratio value for each pixel of that pixel, a normal distribution with a peak value of perfect uniformity is obtained in the case of healthy individuals. In contrast, the distribution of the ratio value is lost in those suffering from diseases. Furthermore, as mentioned above, the method of determining the width using logarithms is merely an example, and the invention is not limited thereto. That is, the invention is a method for implementing "image extraction" as follows: (average density changes over a rough range) ≈ (heartbeat) ≈ (cardiac changes) ≈ (electrocardiogram) ≈ (changes in heart surface area and volume), and is also applicable to methods other than logarithmic methods.

[0156] Furthermore, when considering 3D, by using other devices to measure heart rate, cardiac surface area or volume, cardiac output, and central blood flow, one can measure "local cardiac surface area," "partial cardiac volume," and "blood flow" from these rates in each region. As these quantitative measurements, if cardiac surface area, cardiac volume, cardiac output, and central blood flow can be measured by other means, functional capacity estimates can be estimated from the number of frames, their rates, and the rate of change in that region. That is, in the case of cardiac function analysis, cardiac volume can be estimated based on cardiac motion; in the case of blood flow analysis, pulmonary blood flow can be estimated based on cardiac output, and the estimated blood flow (rate) in bifurcation vessels derived from the central blood flow (rate) can be estimated.

[0157] Furthermore, as mentioned above, measurements can be made more accurately if the entire acquired database can be calculated, but performing computer analysis typically takes time. Therefore, this can be achieved by extracting only a fixed number (phases). For example, automated acquisition can start from the latter half of the image, i.e., from the middle to the back of the image, rather than from the end (manual acquisition is also possible). This makes it easier to extract more stable images by cropping images taken under stress during the first patient photograph. Alternatively, the acquired images can be calculated not as is (e.g., 300 images), but by selecting changes in "heart motion" based on the heart position measured during the first measurement, and then performing the calculation. This allows images that appear to be continuous heartbeats to be labeled when repeated motion image labeling is required. In this case, the calculation can be completed through image labeling. Additionally, when identifying cardiac (myocardial) regions, it is desirable to correct the graphics by using noise trimming and least squares methods, even when manually changing only a portion of the shape or only a portion of the graphic label.

[0158] As described above, according to this embodiment, human images can be evaluated using an X-ray motion imaging device. If digital data is available, calculations can be performed in a generally efficient manner using equipment from existing facilities, thereby reducing installation costs. For example, according to an X-ray motion imaging device equipped with a flat panel detector, the examination of the subject can be completed simply. In cardiac function analysis, screening for myocardial infarction becomes possible. For example, according to an X-ray motion imaging device equipped with a flat panel detector, useless examinations can be eliminated by performing the diagnostic support procedure of this embodiment before CT. Furthermore, due to the simplified examination, urgent diseases can be detected early and treated preferentially. Under current imaging methods, some problems exist in other modalities such as CT and MR, but if these problems can be resolved, detailed diagnosis of each region can be performed.

[0159] Furthermore, it is applicable to the screening of various vascular types, such as cervical blood flow narrowing; and also suitable for the assessment and screening of blood flow in large vessels. Additionally, it can be used to assess characteristic states before and after surgery. Moreover, by performing Fourier transforms on the periodicity of cardiac motion and blood flow to eliminate cardiac and blood flow waveforms, abnormalities in residual biological motion, such as intestinal obstruction, can be observed in abdominal X-ray images.

[0160] Furthermore, when the initially acquired image exhibits a certain high resolution, the number of pixels is large, often requiring significant computation time. In such cases, computation can be performed after the image is reduced to a fixed number of pixels. For example, computation time can be shortened by reducing the image from [4000×4000] pixels to [1028×1028] pixels.

[0161] Furthermore, traditionally, the contrast range of the target image to be determined is manually adjusted; or, a method based on certain standards to relatively extract the target image to be determined is used. However, this is not strictly implemented using frame recognition of the target (e.g., lung field). By detecting the heart region and filtering bone density (cutting out a certain range of low permeability) according to the present invention, the permeability width in the remaining regions is strictly limited to the width of the heart recognition region. This allows for more stringent adjustment of the detectable permeability during the detection of the heart region when reviewed by a person completing the detection, such as a doctor or technician. Furthermore, it allows for more stringent adjustment of permeability suitable for colorimetric evaluation.

[0162] Furthermore, in XP and CY, X-ray transmittance varies depending on the patient's condition to minimize exposure. This transmittance can change during imaging due to lung movement. Similarly, in MRI, the image is captured as a non-uniform signal due to variations in the magnetic field in certain directions. In contrast, by adjusting the calculation of transmittance from the surrounding "background" to a constant form and then back to the transmittance of the specific organ, or by applying a fixed correction to the signal value to correct the non-uniformity of the overall magnetic field variation, changes in the "density / intensity" of a specific organ can be measured more accurately. Moreover, by adjusting the transmittance based on the characteristics of each organ according to the numerical variations in imaging conditions, and by precisely correcting for changes in the "density / intensity" of a specific organ, more accurate amounts of change can be calculated.

[0163] In this invention, average values ​​or intensity changes can be calculated, but data (intensity) obtained from X-rays, CT scans, etc., can undergo "gamma correction," and the density may not be correctly reflected. This gamma correction is a process used to correct the saturation and brightness of images displayed on a monitor. Typically, a computer displays an image on a monitor based on an input signal, but brightness and saturation may vary depending on the monitor's characteristics. Therefore, gamma correction is used to correct these errors. Gamma correction adjusts the relative relationships between signal and color data input and output to the monitor, resulting in a more natural display effect. However, since the gamma-corrected image is no longer the original image, it may cause inconvenience when applying image processing according to the invention. Therefore, by applying "inverse gamma correction," i.e., an inverse filter corresponding to the gamma correction value, to the gamma-corrected image, the image before gamma correction is obtained. This allows obtaining the image before gamma correction, thereby enabling the correct implementation of image processing according to the invention.

[0164] Furthermore, this invention is not limited to gamma correction; it can also be used to restore images that have undergone other image processing methods. For example, by analogy with image processing performed on the image based on pixel value changes in regions with constant density during shooting, such as the external space of the human body, a function that keeps pixel value changes constant is applied to all pixels. This allows for the acquisition of an image close to the original image and enables the correct execution of image processing according to the invention.

[0165] Furthermore, the present invention can recover individual images from overlaid images. Traditionally, "X-ray differential imaging technology" is known. This technology overlays "past and present X-ray images" of the same patient taken during medical examinations, etc., and emphasizes areas that have changed from the past to the present, i.e., areas assumed to be abnormal. For example, cancer can be detected at an early stage by doing so. As a method of overlay, when multiple images are taken as "first, second, third, fourth…", for example, "first, second, third" are used as "overlay images of the first image"; while "second, third, fourth" are used as "overlay images of the second image"; and "third, fourth, fifth" are used as "overlay images of the third image". For such overlaid images, the pixel values ​​are higher when multiple images overlap, and lower when they are not overlaid. Since such overlaid images may blur the outlines of organs, it is desirable to restore them to their respective original images. According to the present invention, the outlines of organs can be identified, making it possible to recover the original images from the overlaid images.

[0166] Furthermore, this invention can also image the frequency differences in each region of an organ. In other words, in an organ undergoing periodic movement, the change values ​​of each region can be Fourier transformed, and each region can be weighted based on the composition ratio of frequency components, such as by color classification, to display the characteristics of each region. For example, the frequency components that reach their peak in each region can be identified and each region can be classified by color. Furthermore, in each region, the percentage of each frequency component within a specific frequency band can be identified, and each region can be classified by color according to that percentage. Moreover, in a specific region, if the standard frequency composition ratio is, for example, "50% at 10Hz, 50% at 20Hz," then the deviation rate from the standard frequency composition ratio can also be visualized. In this way, for example, for the heart, by displaying the spectral distribution of frequencies by color classification, it is easy to determine at a glance whether correct or incorrect movement has occurred.

[0167] Label Explanation

[0168] 1. Basic Module

[0169] 3. Cardiac Function Analysis Unit

[0170] 5. Blood Flow Analysis Unit

[0171] 7. Another blood flow analysis unit

[0172] 9. Fourier Analysis Unit

[0173] 10 Waveform Analysis Unit

[0174] 11 Visualization / Digitalization Unit

[0175] 13 Input Interface

[0176] 15 Databases

[0177] 17 Output Interfaces

[0178] 19. Monitor.

Claims

1. A computer-readable medium storing a diagnostic support program that analyzes images of human organs and displays the analysis results, wherein when executed by a processor, the program causes a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; Calculate a portion of organ-specific periodic motion between each of the frame images based on the pixel values ​​between each of the frame images, or process organ-specific frequencies between each of the frame images; Perform Fourier transform processing on the image changes between each frame; The process of extracting the spectrum within a fixed frequency band from the spectrum obtained after the Fourier transform, wherein the fixed frequency band includes the spectrum corresponding to a portion of the spectrum that shows a specific periodic motion of the organ or a frequency specific to the organ; The spectrum extracted from the fixed frequency band is processed by inverse Fourier transform; and The processing of each image after the inverse Fourier transform is output.

2. The computer-readable medium of claim 1, wherein the process further comprises: The process involves segmenting images of human organs into multiple block regions and calculating image changes in each block region within each frame image; The Fourier transform described therein transforms the changes in the image of each block region in each frame image.

3. The computer-readable medium according to claim 1, further comprising: Based on the pixel values ​​between the frame images, a portion of the waveform showing organ-specific periodic motion is calculated between each of the frame images, or the rate of change of organ-specific frequency is processed between each of the frame images; The processing selects a color corresponding to a portion of a waveform displaying a specific periodic motion of the organ, or the rate of change of a specific frequency of the organ; and The process of displaying an image on a monitor by adding a selected color to the rate of change of pixel values.

4. The computer-readable medium according to claim 3, further comprising: The process involves segmenting an organ image into multiple block regions and calculating the rate of change of the image in each block region within each frame image; in In the process of selecting a color corresponding to a portion of a waveform that displays a specific periodic motion of an organ, or the rate of change of a specific frequency of the organ, a color is selected for each block region.

5. A computer-readable medium storing a diagnostic support program that analyzes images of human organs and displays the analysis results, wherein when executed by a processor, the program causes a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; Calculate a portion of the waveform that displays organ-specific periodic motion between each of the frame images based on the pixel values ​​between each of the frame images, or process organ-specific frequencies between each of the frame images; The processing involves extracting a portion of the waveform corresponding to the periodic motion of an organ in the image through a digital filter, or pixels in the image that vary at a frequency specific to an organ. and The output includes processing of the image of the pixels extracted by the digital filter.

6. The computer-readable medium according to claim 5, further comprising: The process involves segmenting an organ image into multiple block regions and calculating the changes in the image within each block region of the frame image; in In the process of extracting a portion of a waveform corresponding to an organ-specific periodic motion in the image, or pixels that vary at an organ-specific frequency in the image, the digital filter is used to extract the pixels for each block region.

7. The computer-readable medium according to claim 2, 4 or 6, wherein The block region is formed using the von Lonoy partitioning method.

8. The computer-readable medium according to claim 2, 4 or 6, wherein Regardless of whether they are adjacent frames, the changes in the image in the block region of each frame image are calculated while maintaining the relative positional relationship of the pixels indicating the organ between each frame image.

9. The computer-readable medium according to claim 2, 4 or 6, wherein Regardless of whether they are adjacent frames, the rate of change is calculated from the aspects of change in the entire organ between each frame image, and the change of the image in a specific block region is calculated based on the calculated rate of change.

10. The computer-readable medium of claim 9, wherein The rate of change varies depending on the location of the block region within the organ, or it may be constant throughout the organ.

11. A computer-readable medium storing a diagnostic support program that analyzes images of human organs and displays the analysis results, wherein when executed by a processor, the program causes a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; Calculate a portion of the waveform that displays organ-specific periodic motion between each of the frame images based on the pixel values ​​between each of the frame images, or process organ-specific frequencies between each of the frame images; Processing that limits the analysis scope for all acquired frame images; The process of dividing the analysis scope into multiple regions; and The processing of each segmented region is classified based on an index that shows a portion or frequency of a waveform exhibiting periodic motion.

12. The computer-readable medium according to any one of claims 1-6 or 11, wherein The pixel value of each frame image is calculated while maintaining the absolute positional relationship of each pixel in the plurality of frame images.

13. The computer-readable medium of claim 12, wherein... Multiple frames of images are divided into multiple groups, and periodic changes representing organ states are calculated while maintaining the absolute positional relationship of each pixel in the multiple frames of images belonging to each group.

14. The computer-readable medium according to any one of claims 1-6, 11 or 13, further comprising a process for restoring the transmittance of a particular region to a particular shape in a frame image taken by X-rays.

15. The computer-readable medium according to any one of claims 1-6, 11 or 13, further comprising the processing of correcting the signal values ​​of regions of magnetic field inhomogeneity in the frame image taken by MRI and converting them into an image obtained when the magnetic field is uniform.

16. The computer-readable medium of claim 11, wherein The process includes: The process of calculating the difference or ratio between any given frame image and other frame images selected at arbitrary intervals; as well as The calculated difference or ratio is used to visualize the analysis range set for each frame of the image.

17. The computer-readable medium of claim 11, wherein... The process includes: The pixel values ​​within each frame image are calculated as relative or logarithmic values. The process of calculating the ratio of any frame image, represented as the relative value or logarithmic value, to other frame images, represented as the relative value or logarithmic value and selected at arbitrary intervals; as well as The calculated ratio is then visualized.

18. The computer-readable medium of claim 11, wherein... The process includes: The pixel values ​​within each frame image are calculated as relative or logarithmic values. as well as The calculated difference or ratio is used to visualize the analysis range set for each frame image; During the process, the pixel values ​​of the frame image, after excluding artifacts, are calculated.

19. The computer-readable medium of claim 11, wherein The process includes: The pixel values ​​within each frame image are calculated as relative or logarithmic values. The process of calculating the ratio of any frame image, represented as the relative value or logarithmic value, to other frame images, represented as the relative value or logarithmic value and selected at arbitrary intervals; as well as The calculated ratio is then visualized. During the process, the pixel values ​​of the frame image, after excluding artifacts, are calculated.

20. A computer-readable medium having stored thereon a diagnostic support program for analyzing images of human organs and displaying the analysis results, the program, when executed by a processor, causing a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; The process of calculating periodic changes in organ states between the frame images based on pixel values; Perform Fourier transform processing on the image changes between each frame; The process of extracting the spectrum within a fixed frequency band from the spectrum obtained after the Fourier transform, wherein the fixed frequency band includes the spectrum corresponding to the frequency of organ movement; The spectrum extracted from the fixed frequency band is processed by inverse Fourier transform; Output the processing of each image after the inverse Fourier transform; Set the maximum outer edge processing when maximizing organ size; Processing for setting the minimum outer edge when minimizing organ size; The process of calculating the coefficients of the outer edges of organs of other sizes in each image using the maximum outer edge and the minimum outer edge; The processing of the waveform and control points of the waveform, which correspond to the coefficients of the outer edge of the organ in each image, is displayed in the graph. The processing involves changing the coefficients of each image by altering the position of the control points.

21. The computer-readable medium of claim 20, wherein... The average pixel value of the image of the organ is displayed on the graph.

22. The computer-readable medium according to claim 20 or 21, wherein The image of the organ is displayed side-by-side with the graphic.

23. A computer-readable medium having stored thereon a diagnostic support program for analyzing images of human organs and displaying the analysis results, the program, when executed by a processor, causing a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; The process of segmenting the image of a human organ in each frame image into multiple block regions; The process of calculating changes in the image within a block region of each frame image; The value of the change in each of the block regions is processed by Fourier transform; and The process of classifying each region by color based on the composition ratio of frequency components.

24. A computer-readable medium having stored thereon a diagnostic support program for analyzing images of human organs and displaying the analysis results, the program, when executed by a processor, causing a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; Extract pixels with different coordinates from each pixel in a specific frame image and in the next frame image and subsequent frames image, and calculate a portion of a waveform showing organ-specific periodic motion between each of the frames image, or organ-specific frequency processing between each of the frames image; The changes in the images between each frame are processed by Fourier transform; The spectrum obtained after the Fourier transform is processed to extract the spectrum from a fixed frequency band that corresponds to a portion of the waveform showing organ-specific periodicity or the spectrum of organ-specific frequency. The spectrum extracted from the fixed frequency band is processed by inverse Fourier transform. and The processing of each image is output after the inverse Fourier transform is performed.

25. A computer-readable medium having stored thereon a diagnostic support program for analyzing images of human organs and displaying the analysis results, the program, when executed by a processor, causing a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; For each pixel in a specific frame image, extract pixels with coordinates different from those in the next frame image and subsequent frame images and calculate a portion of a waveform that shows organ-specific periodic motion, or process organ-specific frequencies between each of the frame images; The processing involves extracting a portion of the waveform corresponding to the periodic motion of an organ in the image through a digital filter, or pixels in the image that vary at a frequency specific to an organ. and The output includes the processed image of the pixels extracted by the digital filter.

26. A computer-readable medium storing a diagnostic support program that analyzes images of human organs and displays the analysis results, wherein when executed by a processor, the program causes a computer to perform the following process, the process comprising: Processing of acquiring multiple frame images; For each pixel in a specific frame image, extract pixels with coordinates different from those in the next frame image and subsequent frame images and calculate a portion of a waveform that shows organ-specific periodic motion, or process organ-specific frequencies between each of the frame images; Based on a portion of the waveform showing organ-specific periodic motion between each of the frame images, or the processing that defines the analysis range for all the acquired frame images based on organ-specific frequencies between each of the frame images; The process of calculating the rate of change between the frame images within the analysis range; The processing selects the color corresponding to the rate of change between each frame image; and The process of displaying an image on a monitor using selected colors.