The present invention relates to an ultrasonographic device, and particularly to an improvement of an ultrasonographic device having a function of matching tomogram information of an examinee obtained through an ultrasonic probe with diagnosis image information taken from a database and displaying a matching result on a display device to support and assist subsequent measurement, diagnosis of the examinee.
it is very useful for diagnosis of pathologic lesion of heart to quantitatively estimate, for example, a blood flow, the volume of a heart chamber and strain of heart muscle in diagnosis of cardiac functions by using an ultrasonographic device.
Therefore, when the diagnosis of the cardiac functions is performed by using an ultrasonographic device, tomograms of a heart are obtained from various directions through an ultrasonic probe and these tomograms are observed and measured. For example, a valve, the wall thickness of heart muscle, a wall motion can be observed on a short-axis view, and a valve, the volume of a heart chamber, the blood flow, the wall thickness of heart muscle, wall motion can be obtained on a long-axis view.
In a case where the diagnosis of the cardiac functions is performed while a tomogram as described above is observed, if the type of such a view displayed on a display device can be easily identified, for example, if it can be identified whether the view is a short-axis view or a long-axis view or the like, the subsequent measurement and diagnosis are extremely efficient.
Therefore, for example, JP-A-2002-140689 discloses, for example, a medical image processing device used in combination with the ultrasonographic device in which cross-sectional image analyzing means for storing image patterns of respective cross-sectional images of a living organ, particularly a heart as dictionary images compares a two-dimensional cross-sectional image obtained by obtaining means with the stored dictionary images, cross-sectional information comprising the three-dimensional cross-sectional position and cross-sectional direction of the cross-sectional image concerned is analyzed on the basis of a dictionary image having a similar pattern, an analysis result is output to output means, and the cross-sectional position and direction are automatically recognized from the obtained cross-sectional image data of the living body, so that structural information of a target object which will be an aid for diagnosis of a doctor can be extracted.
However, the analysis of the cross-sectional information in a medical image processing device used for the ultrasonographic device disclosed in JP-A-2002-140689 is based on the assumption that an examiner can first obtain a tomogram proper to measurement, and it deals with the analysis of the cross-sectional information after the proper tomogram is obtained.
Therefore, in the cardiac function measurement, the examiner is required to select a measurement item from a large number of measurement menus in accordance with a measurement purpose before a measurement is carried out, and also required to carry out an examination while checking it on the basis of examiner's experience whether a proper view matched with the measurement purpose can be obtained.
An object of the present invention is to provide an ultrasonographic device that can automatically identify the type of a view obtained by an examiner on a real-time basis.
Representatives of the invention disclosed in this application will be briefly described hereunder.
(1) An ultrasonographic device according to the present invention is characterized by comprising, for example, a template conversion unit for converting tomogram information of an examinee to tomogram information which is formed into a template; a database for storing diagnosis image information formed into a template in advance; a matching unit for matching the templated tomogram information of the examinee with the templated diagnosis image information; and a display device for displaying a matching result of the matching unit.
(2) An ultrasonographic device according to the present invention is, for example, based on the construction of (1) and characterized in that when the matching between the templated tomogram image and the templated diagnosis image information is satisfied by the matching unit, a symbol mark such as a body mark or a character which represents the type of the tomogram information is displayed on the display device together with a tomogram image thereof.
(3) An ultrasonographic device according to the present invention is, for example, based on the construction of (1) and characterized in that templated diagnosis image information taken from the database contains not only standard diagnosis image information of an able-bodied person, but also diagnosis image information containing a lesioned part, and when the matching between the templated tomogram information and the diagnosis image information containing the lesioned part is satisfied by the matching unit, a disease name is displayed on the display device together with the tomogram image thereof.
(4) An ultrasonographic device according to the present invention is, for example, based on the construction of (1) and characterized in that tomogram information of an examinee obtained through an ultrasonic probe is tomogram information of a heart which is obtained in synchronization with some time phase of an electrocardiogram of the examinee.
According to the present invention, it can be checked on a real-time basis that a proper view matched with a measurement purpose can be obtained, so that the precision of the measurement based on the check concerned is enhanced, and the efficiency of a subsequent examining diagnosis is drastically enhanced.
An embodiment of an ultrasonographic device according to the present invention will be described hereunder with reference to the drawing.
Diagnosis image information is stored into the database with being successively passed through an input device 7 containing a mouse, a keyboard, a track ball and a template data reading unit from a disk or the like, and tomogram data appending means 8 for adding relevant information. Here, the diagnosis image information stored in the database 6 is constructed as so-called templated diagnosis image information formed. As descried later, the diagnosis image information is correlated with tomogram information of an examinee obtained through an ultrasonic probe 1 on a real-time basis. The templated diagnosis information is input not only from the input device 7, but also from a template converting unit 4 described later which is equipped in the ultrasonographic device as occasion demands. This is because the latest templated information which is as much as possible is stored in the database 6. As a method of creating a template is used as an eigen space method, a subspace method, a mutual subspace method or the like.
Furthermore, with respect to each diagnosis image information, the tomogram data appending unit 8 appends other data such as the type of the diagnosis image information, etc. in association with the diagnosis image information. Accordingly, when one of the diagnosis image information is selected, various kinds of data associated with the diagnosis image information, for example, the type name of the image, the age, sex and disease name of an examinee, a standard measurement initial setting value of the image and measurement set values which were past set by the user can be obtained.
A set which is matched with the measurement and the diagnosis purpose is prepared as an electronic file for the database 6. The content of the set matched with the purpose is an assembly containing the parasternal view (left ventricular long-axis view, left ventricular short-axis view, right ventricular inflow path long-axis cross-section) or the cross-sections of an apical site approach image (left ventricular long-axis cross-section, four-chamber cross-section, five-chamber cross-section, two-chamber cross-section) which are obtained by a normal examination for the purpose of identifying the type of the cross-section, an assembly containing tomograms of hypertrophic cardiomyopathy (containing various hypertropic styles), dilated cardiomyopathy, etc. which are obtained for the purpose of identifying the type of cardiomyopathy, an assembly containing tomograms of mitral valve, aortic valve disease, tricuspid valve disease, etc. which are obtained for the purpose of identifying the type of valve disease, and an assembly containing tomograms of the interatrial septum defect, the interventricular septum defect, etc. which are obtained for the purpose of identifying the type of congenital cardiac disease, for example.
These sets may be used independently of one another, or these sets may be integrated into one set so that identification of the view and identification of the type are performed at the same time. Furthermore, various kinds of templated diagnosis image information which correspond to the sex and age of a patient can be prepared every display depth or view, whereby a database to identify an input tomogram of a broad range can be prepared. If the set of the database is set so as to contain some specific disease data, it can be used to identify a disease. For example, if the database of dilated cardiomyopathy and hypertrophic cardiomyopathy is prepared, it is effective to identify a lesion or a disease from tomograms obtained when the shape of a heart is different from a normal one. Furthermore, the user can create a new database set based on the viewpoint at the user side, and add/delete data to/from the database set, thereby constructing a database corresponding to the use situation of the user.
The ultrasonographic device is equipped with an ultrasonic probe 1 used in contact with the body surface of the examinee or the like, and a signal value storage unit 2 for storing tomogram information (signal) of the examinee which is created by a reflection echo signal in the examinee obtained through the ultrasonic probe 1. With respect to the tomogram information, as is apparent from the description made later with reference to
The ultrasonographic device is equipped with a signal value extracting unit 3 for extracting prescribed tomogram information from the tomogram information (signal) stored in the signal value storage unit 2. The tomogram information to be extracted may be the directly extracted amplitude value of the information, or extracted by calculating a static amount in any area. For example, when a sector type ultrasonic probe is used as the ultrasonic probe 1, all echo signals of a tomogram plane DL constituting a fan, or the brightness value of an image may be extracted. Furthermore, signals at respective positions which are scanned in fan-like fashion on the tomogram plane DL may be extracted as shown in
When the signals at the respective positions scanned in a fan-like fashion on the tomogram plane DL as shown in
The tomogram information extracted by this signal value extracting unit 3 is converted to templated tomogram information by the template conversion unit 4. By creating the templated tomogram information as described above, the matching with the corresponding templated diagnosis image information from the database can be easily performed in an extremely short time.
As a templating method, when the amount of data is reduced or a feature portion is extracted, the eigen space method as described above may be used, and when the signal values of plural time phases are obtained, the subspace method may be used. The tomogram information which is templated according to the predetermined method as described above is templated is templated in a predetermined registerable form, and stored in the database 6 as occasion demands.
Here, the method of obtaining tomogram information templated in a registerable form in the database 6 by using the eigen space method will be described.
First, a set of signal value vectors of tomogram information obtained from the signal value extracting unit 3 is represented by A. The number of tomograms used for templating is represented by N, and the average value (A) and covariance matrix S thereof are calculated according to the following equation to create the eigen space. T represents transposition.
S is subjected to eigen value decomposition to calculate eigen values λ and eigenvectors u. The eigen values λ are arranged in descending order, and an eigenvector Φ={u1, u2, . . . , uk} corresponding to any k is determined from the largest eigen value, and a database is complied by setting as a template a vector A* obtained by projecting A in the following equation (3). Accordingly, the capacity of the database is reduced, and also a database of only the feature serving as a main part of the signal can be efficiently made.
A*=Φ
T(A−Ā) (3)
Furthermore, a method using the subspace method is known as another method of templating so as to be registerable in the database 6. By using signal values of L frames (for example, selected from P, Q, R, S, T, U waves on ECG or another frame) with respect to a signal value vector Ap of some cross-section p, the average value (
Sp is subjected to eigen value decomposition to calculate eigen values λp and eigenvectors up. The eigen values are arranged in descending order, and an eigenvector Φp={up1, up2, . . . , upk} corresponding to any k is determined from the largest eigen value, and a database is complied by setting as a template a vector A*p obtained by projecting A in the following equation (6). Accordingly, the capacity of the database is reduced, and also a database of only the feature serving as a main part of the image can be efficiently made. Furthermore, the signal varies at all times, and thus a feature which cannot be obtained by only a single signal can be stably extracted by using plural signals.
A*p=Φp
T(A−Āp) (6)
A signal value vector Anew in which the type of the cross-section input from the template conversion unit 4 for new matching is unknown, and the database in which cross-section types are known are subjected to the matching processing in a matching unit 5. In this processing, the cross-section having the nearest feature to the feature of the signal value vector Anew of the input cross-section is found out from the diagnosis image information of the database 6.
When the database 6 templated by using the eigen space method, Anew is projected to the eigen space as indicated in the following equation (7).
A*
new=ΦT(Anew−Ā) (7)
The difference between the projected vector A*new and A* of the equation (3) is quantified, and the view having the smallest difference is found out. The type of the view having the smallest difference is set as the identified view type. A scale such as Euclidean distance, Mahalanobis distance, Manhattan distance, degree of similarity or the like may be used for the calculation of this difference. The above scale is used when the difference is quantified.
When the database 6 templated by using the subspace method is used, Anew is projected to the subspace of each view p as indicated by the equation (8).
A*
pnew=ΦTp(Anew−Āp) (8)
The difference between the projected vector A*pnew and A*p of the equation (6) is quantified to identify the view.
In both the methods, a threshold value is provided with respect to the difference between the templated input and the template in the database, and if the difference is larger than the threshold value, the matching is set to be impossible.
Furthermore, the mutual subspace method of inputting plural image signals, creating the templated input subspace and matching the input subspace and the subspace of the database may be used.
As described above, the templated tomogram of the examinee obtained through the ultrasonic probe 1 is correlated to the templated diagnosis image information taken from the database 6 in the matching unit 5, and when the difference therebetween is not more than a predetermined threshold value, a body mark 12 as a symbol mark representing the type of the tomogram information and a character 13 are displayed on the display plane 9A of the display device 9 as shown in
In this case, when diagnosis image information containing disease data is prepared as diagnosis image information stored in the database 6, a disease may be displayed on the display plane 9A. Furthermore, when it is impossible to correlate the tomogram image obtained through the ultrasonic probe 1 with the diagnosis image information taken from the database 6, a mark indicating “matching-impossible” may be displayed or non-display may be made. As the display of the body mark 12 and the character 13, the matching result may be renewed every R-wave is detected, or may be renewed every time the matching processing is completed. Accordingly, the examiner can check whether the tomogram matched with the examination purpose can be correctly obtained.
A measurement setting unit 10 renews the measurement menu and the initial set value on the basis of the identified view type. For example, when a short-axis view is identified, measurement items used for only the long-axis view are deleted, and a menu is constructed by only measurement items used for the short-axis view. When the examiner enters the measurement mode, only the short-axis menu is displayed. Accordingly, the load imposed on the cumbersome measurement item selection from all the measurement items like the prior art can be reduced. The same operation is also carried out when another view is identified. Furthermore, measurement set values which are made in the form of a database as a set with the identified views may be called and automatically input as the initial set values.
When the set value of the measurement item is set in the measurement set unit 10, a measurement is executed on the tomogram displayed on the basis of the set value in the measuring unit 11, and the result thereof is displayed on the display device 9.
In this embodiment, the templated diagnosis image information stored in the database 6 contains information created at some time phase of an electrocardiogram (ECG), for example, in synchronization with R-wave, and in the matching described later, information created in synchronization with the corresponding time phase is read out. The diagnosis image information stored in the database 6 may be only information which is created at some time-phase, for example, in synchronization with the R-wave. When only templates at some predetermined time phase are stored in the database 6, the storage capacity of the database 6 can be reduced as compared with a style in which moving pictures at all time phases are stored, and the reading time is shorter, so that the matching time can be also shortened.
First, R-wave is detected from the examinee in step S1. A reflection echo signal for creating a tomogram is taken in synchronization with the R-wave. In step S2, it is judged whether the identification processing of a previous tomogram is finished or not. The identification processing is required to be finished during one beat of the heart. However, the length of one beat is different among individuals, and if the identification processing is not finished during one beat, the identification processing may be extended until next R-wave is detected. This is because the tomogram is necessarily created in synchronization with the R-wave. If the previous identification processing is finished in step S2, the reflection echo signal is obtained in step S3. In step S4, tomogram information which is templated in such a form that it can be correlated with templates in the database 6 is obtained from the obtained signal.
Then, in step S5, the templated diagnosis image information is taken from the database 6, and correlated with the templated tomogram information. In step S6, it is judged whether the tomogram information is coincident with the diagnosis image information, that is, it is judged whether the tomogram information is identified or not. If it is not identified, the processing waits until next R-wave detection, and the above operation is repeated. If the tomogram image information is identified in step S6, the identification result is displayed on the display device 9 in step S7.
Subsequently, it is judged whether the measurement setting is carried out through the measurement setting unit 10 in accordance with the type of the identified tomogram by the examiner in step S8. If it has not yet been identified, the above operation is repeated. If the measurement setting is carried out, the measurement of the tomogram displayed on the display device is executed according to the setting in step S9, and the measurement result is displayed on the display device 9 in step S10.
In the above embodiment, the electrocardiogram (ECG) based on the electrocardiograph can be prepared. However, for example, when the electrocardiogram (ECG) cannot be immediately prepared like an emergency case or the like, the time phase on ECG is estimated from the motion of the heart. The moving speed of the heart is not fixed during one beat. Accordingly, it is possible that the moving speed is measured, the time phase at a specific speed, for example, at the highest speed or the lowest speed is detected and the signal is taken in synchronization with the time phase. In this case, in order to measure the moving speed of the heart, a method called as optical flow may be used, for example. That is, as shown in
Number | Date | Country | Kind |
---|---|---|---|
2005-330523 | Nov 2005 | JP | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/JP2006/322727 | 11/15/2006 | WO | 00 | 7/15/2009 |