System and method for an otitis media database constructing and an image analyzing

Abstract

The invention provides a system and a method for an otitis media database constructing. The steps of method comprise: First, receiving a plurality of tympanic membrane images, wherein the tympanic membrane images are ear infection in different types. Second, choosing one of the tympanic membrane images, then classifying into a plurality of anatomic regions based on a plurality of tissue types. Coding each of anatomic regions with a numerical number to describe its morbid condition, and obtaining an eigenvalue through collecting the numerical number of each anatomic region. Furthermore, repeatedly choosing other one of tympanic membrane images until each tympanic membrane image obtaining the eigenvalue. And obtaining a matrix through collecting the eigenvalue of each tympanic membrane image, then generating an otitis media database.

Claims

1. A tympanic membrane image analysis method, comprising: receiving a tympanic membrane image obtained from an otoscope, and classifying the tympanic membrane image into a plurality of anatomic regions based on a plurality of tissue types; coding each of the anatomic regions with a numerical number to describe its morbid condition; obtaining an eigenvalue through collecting the numerical numbers from each anatomic region; obtaining a matrix with a plurality of reference eigenvalues from an otitis media database, then calculating a plurality of contrast values by performing multiplication of the matrix and the eigenvalue with a weighting matrix for the numerical numbers of each anatomic region; and determining a type of ear infection from the tympanic membrane image in accordance with the lowest contrast value, the lowest contrast value being a reference eigenvalue; wherein the otitis media database is constructed through steps comprising: receiving a plurality of reference tympanic membrane images, wherein the reference tympanic membrane images are ear infections in different types; choosing one of the reference tympanic membrane images, and classifying into the anatomic regions based on the tissue types; coding each of the anatomic regions with the numerical number to describe its morbid condition; obtaining the reference eigenvalue through collecting the numerical number of each anatomic region; choosing other one of the reference tympanic membrane images repeatedly until obtaining the reference eigenvalues from each tympanic membrane image; and determining the otitis media database of ear infection in accordance with the matrix through sequentially arranging the reference eigenvalues of each reference tympanic membrane image.

2. The method of claim 1, wherein the ear infection comprises an acute otitis media and an otitis media with effusion.

3. The method of claim 1, wherein the morbid condition of the anatomic region comprises at least one of: color feature, geometric feature, texture feature and shape feature.

4. The method of claim 3, wherein the color feature comprises at least one of: hue, saturation and lightness.

5. The method of claim 3, wherein the geometric feature comprises a histogram of oriented gradient (HOG).

6. The method of claim 3, wherein the texture feature comprises a local binary pattern (LBP).

7. The method of claim 3, wherein the shape feature comprises a statistical chart of self-similarity geometric pattern.

8. A tympanic membrane image analysis system, comprises: an access device with an otitis media database, saving a matrix; an otoscope, receiving a tympanic membrane image; a computation processor connected with the otoscope, classifying the tympanic membrane image into a plurality of anatomic regions based on a plurality of tissue types, then coding each of anatomic regions with a numerical number to describe its morbid condition, and obtaining an eigenvalue through collecting the numerical numbers from each anatomic region; and an output device connected with the computation processor and the access device, obtaining a matrix from the otitis media database, then calculating a plurality of contrast values by performing multiplication of the matrix and the eigenvalue with a weighting matrix for the numerical numbers of each anatomic region, determining a type of ear infection from the tympanic membrane image in accordance with the lowest contrast value, the lowest contrast value being a reference eigenvalue; wherein the otitis media database is constructed through steps comprising: receiving a plurality of reference tympanic membrane images, wherein the reference tympanic membrane images are ear infections in different types; choosing one of the reference tympanic membrane images, and classifying into the anatomic regions based on the tissue types; coding each of the anatomic regions with the numerical number to describe its morbid condition; obtaining the reference eigenvalue through collecting the numerical number of each anatomic region; choosing other one of the reference tympanic membrane images repeatedly until obtaining the reference eigenvalues from each tympanic membrane image; and determining the otitis media database of ear infection in accordance with the matrix through sequentially arranging the reference eigenvalues of each reference tympanic membrane image.

9. The device according to claim 8, wherein the ear infection comprises an acute otitis media and an otitis media with effusion.

10. The device according to claim 8, wherein the morbid condition of the anatomic region comprises at least one of: color feature, geometric feature, texture feature and shape feature.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Many aspects of the disclosure can be better understood with reference to the following drawings. The components in the drawings are not necessarily to scale, emphasis instead being placed upon clearly illustrating the principles of the present disclosure. Moreover, in the drawings, like reference numerals designate corresponding parts throughout the several views.

(2) FIG. 1 is a block diagram of a system for database constructing in accordance with various embodiments of the present disclosure.

(3) FIGS. 2A and 2B illustrate diagram of the eigenvalue classification in accordance with various embodiments of the present disclosure.

(4) FIG. 3 is a block diagram of an image analysis system in accordance with various embodiments of the present disclosure.

(5) FIG. 4 illustrates calculation of the lowest difference value in accordance with various embodiments of the present disclosure.

(6) FIG. 5 is a flowchart in accordance with one embodiment for database constructing.

(7) FIG. 6 is a flowchart in accordance with one embodiment for an image analysis.

DETAILED DESCRIPTION OF THE INVENTION

(8) For clarity of disclosure, and not by way of limitation, the detailed description of the invention is divided into the subsections that follow. The present invention provides merely an example of the different types of functional arrangements that may be employed to implement the operation in the various components of a system, such as a computer system connected to an otoscope, a video-otoscope, a wireless video otoscope, and so forth.

(9) The execution steps of the present invention may include application specific software which may store in any portion or component of the memory including, such as random access memory (RAM), read-only memory (ROM), hard drive, solid-state drive, magneto optical (MO), IC chip, USB flash drive, memory card, optical disc such as compact disc (CD) or digital versatile disc (DVD), floppy disk, ZIP, magnetic tape, or other memory components.

(10) Generally speaking, the method of invention implemented in the computing device may comprise any one of a wide variety of wired and/or wireless computing devices, such as a desktop computer, portable computer, dedicated server computer, multiprocessor computing device, cellular telephone, personal digital assistant (PDA), handheld or pen based computer, embedded appliance, or other devices with Input/output interfaces, and so forth.

(11) FIG. 1 is a block diagram of a system for database constructing in accordance with various embodiments of the present disclosure. The system for database constructing 100 comprises a receiving device 102, a computation processor 104 and an access device 106.

(12) The receiving device 102 is coupled via a wired or wireless communication path to an otoscope 120 with an image capture device for retrieving a plurality of tympanic membrane images, wherein the tympanic membrane images are ear infection in different types. The ear infection may include, for example, an acute otitis media and an otitis media with effusion.

(13) The computation processor 104 is connected with the receiving device 102 to choose one of the tympanic membrane images, and classifies the tympanic membrane image into a plurality of anatomic regions based on a plurality of tissue types. The tympanic membrane has an ectoderm and an endoderm aspect. For example, the computation processor 104 classifies the tympanic membrane image into a first anatomic region and a second anatomic region, wherein the first anatomic region is the endoderm and the second anatomic region is the ectoderm.

(14) In an embodiment, the computation processor 104 can include any custom-made or commercially available processor, a central processing unit (CPU), a semiconductor based microprocessor (in the form of a microchip), a macroprocessor, one or more application specific integrated circuits (ASICs), a plurality of suitably configured digital logic gates, and other well known electrical configurations comprising discrete elements both individually and in various combinations to coordinate the overall operation of the computing system.

(15) Each of anatomic regions is coded with a numerical number (v.sub.11) to describe its morbid condition. The morbid condition of the anatomic region comprises at least one of color feature, geometric feature, texture feature and shape feature. The color feature comprises at least one of hue, saturation and lightness. The geometric feature comprises a histogram of oriented gradient (HOG). The texture feature comprises a local binary pattern (LBP). The shape feature comprises a statistical chart of self-similarity geometric pattern.

(16) Furthermore, the computation processor 104 chooses other one of anatomic regions orderly until all anatomic regions are coded with numerical numbers (v.sub.11, . . . , v.sub.1k), then obtains an eigenvalue (A.sub.1=[v.sub.11, . . . , v.sub.1k]) through collecting the numerical number of each anatomic region.

(17) The computation processor 104 further chooses other one of tympanic membrane images repeatedly until each tympanic membrane image obtaining the eigenvalue (A.sub.2=[v.sub.21, . . . , v.sub.2k], . . . , A.sub.n).

(18) FIGS. 2A and 2B illustrate diagram of the eigenvalue classification in accordance with various embodiments of the present disclosure. FIGS. 2A and 2B show otitis media classification 200 and the definition of the numerical number of each anatomic region 220.

(19) In an embodiment, the anatomic region A is in the tympanic membrane image of acute otitis media, wherein the morbid condition of the anatomic region A is red-pink and the anatomic region A is represented by 2.

(20) The anatomic region B is in the tympanic membrane image of acute otitis media, wherein the morbid condition of the anatomic region B is mild and the anatomic region B is represented by 1.

(21) The anatomic region C is in the tympanic membrane image of acute otitis media, wherein the morbid condition of the anatomic region C is near total and the anatomic region C is represented by 3.

(22) The computation processor 104 identifies a numerical number to each anatomic region until each anatomic region obtaining the numerical number, then obtains an eigenvalue [2 1 3 . . . ] through collecting the numerical number of each anatomic region.

(23) In an embodiment, the anatomic region A is in the tympanic membrane image of otitis media with effusion, wherein the morbid condition of the anatomic region A is amber and the anatomic region A is represented by 1.

(24) The anatomic region B is in the tympanic membrane image of otitis media with effusion, wherein the morbid condition of the anatomic region B is moderate and the anatomic region B is represented by 2.

(25) The anatomic region C is in the tympanic membrane image of otitis media with effusion, wherein the morbid condition of the anatomic region C is mild and the anatomic region C is represented by 1.

(26) The computation processor 104 identifies a numerical number to each anatomic region until each anatomic region obtaining the numerical number, then obtains an eigenvalue [1 2 1 . . . ] through collecting the numerical number of each anatomic region. The computation processor 104 obtains a matrix [2 1 3 . . . ; 1 2 1 . . . ; . . . ] through collecting the eigenvalue of each tympanic membrane image.

(27) The access device 160 is connected with the computation processor 104 to obtain a matrix (A=[v.sub.11, . . . , v.sub.1k v.sub.21, . . . , v.sub.2k, . . . , v.sub.n1, . . . , v.sub.nk]) through collecting the eigenvalue of each tympanic membrane image, then generates an otitis media database 140, wherein n represents the number of tympanic membrane images, and k represents the number of features regions.

(28) In an embodiment, the access device 106 can include any one of a combination of volatile memory elements (e.g., random-access memory (RAM, such as DRAM, and SRAM, etc.)), nonvolatile memory elements (e.g., ROM, hard drive, tape, CDROM, etc.) and other common digital signals storage element.

(29) FIG. 3 is a block diagram of an image analysis system in accordance with various embodiments of the present disclosure. The image analysis system 300 comprises an otitis media database 302, a receiving device 304, a computation processor 306 and an output device 308.

(30) In an embodiment, the otitis media database 302, for example, an optical disc device, a hard drive or a remote server is communicatively via a network. The network, which may include, for example, the Internet, intranets, extranets, wide area networks (WANs), local area networks (LANs), wired networks, wireless networks, or other suitable networks, etc., or any combination of two or more such networks.

(31) The computation processor 306 is connected with the receiving device 304 to classify the tympanic membrane image into a plurality of anatomic regions based on a plurality of tissue types. Then, each of anatomic regions is coded with a numerical number to describe its morbid condition.

(32) In an embodiment, the receiving device 304 is coupled via a wired or wireless communication path to an otoscope 320 with an image capture device for retrieving a plurality of tympanic membrane images.

(33) Then, the computation processor 306 chooses other one of anatomic regions orderly until all anatomic regions are coded with numerical number (v11, . . . , v1k), and obtains an eigenvalue (A1=[v11, . . . , v1k]) through collecting the numerical number of each anatomic region.

(34) The output device 308 is connected with the computation processor 306 and the otitis media database 302, obtains a matrix A=[v.sub.11, . . . , v.sub.1k v.sub.21, . . . , v.sub.2k, . . . , v.sub.n1, . . . , v.sub.nk] from the otitis media database 302, and then obtains a plurality of contrast values based on performing multiplication of the matrix and a weighting matrix (S=[s.sub.1, s.sub.2, . . . , s.sub.j]).

(35) In an embodiment, the output device 308 is coupled via a wired or wireless communication path to the remote server 340 for recording data from the image analysis system or a display 360, wherein the display 360 may comprise a computer monitor, a plasma screen for a PC, a liquid crystal display (LCD), a touch screen display, or other display device for displaying a table or diagram as results of the image analysis system.

(36) The output device 308 selects a lowest difference value from the difference between the numerical numbers and the contrast values and records a type of ear infection based on the tympanic membrane image corresponding to the lowest difference value of the contrast value.

(37) In an embodiment, the tympanic membrane image has the lowest difference value based on the numerical number minus the contrast value (A.sub.1S), then the output device 308 records the contrast value (A.sub.1S) and the type of ear infection corresponding to the contrast value (A.sub.1S) such as acute otitis media.

(38) Reference is made to FIG. 4, FIG. 4 illustrates calculation of the lowest difference value in accordance with various embodiments of the present disclosure. The output device 308 selects the lowest difference value from the difference between the numerical numbers and the contrast values. The linear combination of the corresponding training features is calculated according to the following expression:

(39) $y^k=\underset{j=1}{\overset{J}{.Math.}}{A}_j^k{s}_j^k+{\mathrm{.Math.}}^k$
where k represents the types of ear infection, y.sup.k represents each feature of the tympanic membrane image, A.sub.j.sup.k represents the matrix, s.sub.j.sup.k represents the weighting matrix, and .sup.k represents a difference value based on performing multiplication of the matrix and the weighting matrix and then subtracted by the numerical number (residual term).

(40) FIG. 5 is a flowchart in accordance with one embodiment for database constructing. It is understood that the flowchart of FIG. 5 provides merely an example of the many different types of functional arrangements that may be employed to implement the operation of the various components of the system for database constructing 100 (FIG. 1). As an alternative, the flowchart of FIG. 5 may be viewed as depicting an example of steps of a method implemented in the system for database constructing 100 according to one or more embodiments.

(41) Beginning with step S500, the receiving device 102 (FIG. 1) retrieves a plurality of tympanic membrane images, wherein the tympanic membrane images are ear infection in different types. The ear infection may include, for example, an acute otitis media and an otitis media with effusion.

(42) In step S502, the computation processor 104 (FIG. 1) is connected with the receiving device 102 to choose one of the tympanic membrane images, and then classifies the tympanic membrane image into a plurality of anatomic regions based on a plurality of tissue types.

(43) In step S504, coding each of anatomic regions with a numerical number to describe its morbid condition. The morbid condition of the anatomic region comprises at least one of color feature, geometric feature, texture feature and shape feature. The color feature comprises at least one of hue, saturation and lightness. The geometric feature comprises a histogram of oriented gradient. The texture feature comprises a local binary pattern. The shape feature comprises a statistical chart of self-similarity geometric pattern.

(44) Then, in step S506, the computation processor 104 chooses other one of anatomic regions orderly until all anatomic regions are coded with the numerical number.

(45) In step S508, the computation processor 104 obtains an eigenvalue through collecting the numerical number of each anatomic region.

(46) In step S510, the computation processor 104 chooses other one of tympanic membrane images repeatedly until each tympanic membrane image obtaining the eigenvalue.

(47) Finally, in step S512, the access device 160 (FIG. 1) is connected with the computation processor 104 to obtain a matrix through collecting the eigenvalue of each tympanic membrane image, and then generating an otitis media database 140.

(48) Reference is made to FIG. 6, which is a flowchart in accordance with one embodiment for the image analysis system. It is understood that the flowchart of FIG. 6 provides merely an example of the many different types of functional arrangements that may be employed to implement the operation of the various components of the image analysis system 300 (FIG. 3). As an alternative, the flowchart of FIG. 6 may be viewed as depicting an example of steps of a method implemented in the image analysis system 300 according to one or more embodiments.

(49) Beginning with step S600, the receiving device 304 (FIG. 3) retrieves a tympanic membrane image.

(50) In step S602, the computation processor 306 (FIG. 3) is connected with the receiving device 304 to classify the tympanic membrane image into a plurality of anatomic regions based on a plurality of tissue types.

(51) In step S604, coding each of anatomic regions with a numerical number to describe its morbid condition. The morbid condition of the anatomic region comprises at least one of color feature, geometric feature, texture feature and shape feature. The color feature comprises at least one of hue, saturation and lightness. The geometric feature comprises a histogram of oriented gradient. The texture feature comprises a local binary pattern. The shape feature comprises a statistical chart of self-similarity geometric pattern.

(52) In step S606, the computation processor 306 further chooses other one of anatomic regions orderly until all anatomic regions are coded with numerical numbers, then obtains an eigenvalue through collecting the numerical number of each anatomic region.

(53) Then, in step 608, the output device 308 (FIG. 1) is connected with the computation processor 306 and the otitis media database 302 for obtaining a matrix from the otitis media database 302, and then obtains a plurality of contrast values based on performing multiplication of the matrix and a weighting matrix.

(54) Finally, in step S610, the output device 308 selects a lowest difference value from the difference between the numerical numbers and the contrast values and recording a type of ear infection based on the tympanic membrane image corresponding to the lowest difference value of the contrast value.

(55) The invention provides a method and system for database constructing based on a plurality of tympanic membrane images of any otitis media. The method and system collects the numerical value of each tympanic membrane image to obtain a matrix, and then generates an otitis media database. The otitis media database comprises a plurality of eigenvalues, and each eigenvalue corresponding to a type of otitis media.

(56) Furthermore, the medical person compares the tympanic membrane images with the otitis media database to diagnosis the tympanic membrane image corresponding to belong to any type of otitis media, not only to accelerate otitis media type of interpretation, but also to improve its accuracy.