KR101484051B1 - Apparatus and method for preprocessing for CAD system using active contour method - Google Patents

Abstract

A preprocessing apparatus for a CAD system using an active contour method is disclosed. The artifact removal unit obtains a binary image using the threshold for the input image, divides the binary image into a predetermined number, calculates a pixel sum for the divided area, and determines the starting point of the pointer And the unnecessary artifacts are retrieved and removed from the input image using the starting point. The contour extractor obtains an outline including the ribbed boundary information, defines an energy function based on the new region, moves gradually inward based on the outline of the input image, and moves when it meets a predetermined reference point One or more outlines are created by connecting the stopped and stopped points. The pectoralis removal unit removes the pectoral muscle by extracting a binary image of the pectoral muscle using the parameters for the mask term of the contour. According to the present invention, different types of artifacts present in a digital mammogram image can be removed for accurate detection of tumor tissue. This can provide a tool for effectively analyzing medical images.

Description

TECHNICAL FIELD The present invention relates to a preprocessing apparatus and method for a CAD system using an active contour method,

The present invention relates to a preprocessing apparatus and method for a CAD system using an active contour method, and more particularly, to a preprocessing algorithm for a mammographic CAD system that removes labels and scanned artifacts and stomach muscles using an active contour method .

Analysis of x-rays, tomography, and region-of-interest (ROI) in MRI is an essential process in computer-aided diagnosis (CAD). For this purpose, homogeneity maps of medical images provide information on analysis of area of interest. Therefore, a lot of research has been conducted over a long period of time for image segmentation in order to detect a region of interest in a medical image.

As one of the researches in the field mentioned above, there is a method based on a partial differential equation (PDE) known as an active contour model for extracting objects of interest in an image. That is, the active contour model is one of the techniques used for image segmentation, in which the curve of the contour proceeds toward the target object boundary, which stops when it encounters the boundary. Therefore, various active contour models and upgraded versions of active contour models have been applied in image processing as well as medical images. However, the conventional technique requires a re-initialization process to set a reference level every time the generation of a homogeneous line, so that it takes a long time and a large amount of computation.

In detail, a method for automatic segmentation of images is disclosed in U.S. Publication No. 2010-0215238 (entitled Method for automatic segmentation of images). This reconstructs the image data for a heart image to distinguish a heart chamber that includes a wall with an inner and outer surface to create an inner contour pointing to the inner surface of the chamber wall to create a pseudo-polar map to determine an outer contour line that points to the outer surface of the chamber, thereby partitioning the reconstructed image using inner and outer contour lines.

In addition, US-A-2010-0226542 discloses a method for generating an image map from an image data set in a method of revealing more meaningful anatomical features of diseased tissues. This distinguishes the boundary surface within a 3D image data set and creates a reference surface from the boundary surface to distinguish the second surface from the 3D image data set to calculate the distance between the point on the reference surface and the point on the second surface, Create an image map based on this.

SUMMARY OF THE INVENTION The present invention is directed to a CAD system using an active contour method for removing different types of artifacts present in a digital mammogram image for accurate detection of tumor tissue. And to provide a preprocessing apparatus and method for the same.

Another object of the present invention is to provide a method and apparatus for detecting a tumor tissue by using an active contour method that removes different kinds of artifacts existing in a digital mammogram image, There is provided a computer-readable recording medium recording a program for causing a computer to execute a preprocessing method for a system.

According to another aspect of the present invention, there is provided a preprocessing apparatus for a CAD system using an active contour method, the apparatus comprising: a binary image obtaining unit that obtains a binary image using a threshold for an input image, An artefact removal unit for dividing an image into a predetermined number and determining a starting point of a pointer by calculating a pixel sum with respect to the divided area, and searching for and removing unnecessary artifacts from the input image using the starting point; And an energy function is defined by acquiring an outline including the rib boundary information, and gradually moves to the inside based on the outermost line of the input image, stops moving when the point corresponding to the predetermined reference is encountered, A contour extractor for concatenating the contour lines to generate at least one contour; And a sternum removing section for extracting a binary image of the pectoral muscle using a parameter for a mask term of the outline to remove the sternum.

According to another aspect of the present invention, there is provided a preprocessing method for a CAD system using an active contour method, the preprocessing method being performed by a preprocessing apparatus for a CAD system using an active contour method, the method comprising the steps of: A binary image is obtained using a threshold, a binary sum is divided into a predetermined number, and a pixel sum is calculated for a divided area to determine a starting point of the pointer, An artifact removal step of searching for and removing an unnecessary artifact to the input image; (b) defining an energy function by acquiring an outline including the ribbed boundary information, moving gradually toward the inside based on the outermost line of the input image, stopping the movement when a point corresponding to a predetermined reference is encountered, An outline extraction step of connecting the stopped points to generate at least one outline; And (c) extracting a binary image of the pectoral muscle using a parameter for the mask term of the contour line to remove the sternal muscle.

According to another aspect of the present invention, there is provided a computer readable medium storing a program for causing a computer to execute any one of the above methods.

According to the preprocessing apparatus and method for a CAD system using the active contour method according to the present invention, in order to accurately detect a tumor tissue, artifacts of different kinds existing in a digital mammogram image Can be removed. This can provide a tool for effectively analyzing medical images.

1 shows a general mammogram image,
FIG. 2 is a block diagram showing the construction of a preprocessing apparatus for a CAD system using the active contour method according to the present invention.
FIG. 3 is a flowchart illustrating a process of removing artifacts according to an embodiment of the present invention.
FIG. 4 is a view showing a polyphase division process using an active contour method;
FIG. 5 illustrates an active contour splitting process according to an embodiment of the present invention.
FIG. 6 is a flowchart illustrating a process of removing a stomach muscle according to an embodiment of the present invention.
FIG. 7 is a flowchart illustrating a preprocessing method for a CAD system using the active contour method according to the present invention.
FIG. 8 is a diagram illustrating a comparison between a pre-processing method for a CAD system using the active contour method according to the present invention,
9 to 10 are diagrams showing divided result images using a preprocessing method for a CAD system using the active contour method according to the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of a pre-processing apparatus and method for a CAD system using an active contour method according to the present invention will be described in detail with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a diagram showing a mammogram image including artifacts and pectoral muscles. Referring to FIG. 1, a digital mammogram image may have artifacts of different kinds. Artifacts may include high intensity labels, low intensity labels, taping or scanning artifacts, and pectoral muscles. If the artifacts are included in the mammogram image, a problem may occur in the detection process and the accuracy of ROI (Region Of Interest) detection may be affected. Therefore, a preprocessing algorithm that can remove these artifacts is important.

FIG. 2 is a block diagram showing a configuration of a preprocessing apparatus 100 for a CAD system using an active contour method according to the present invention.

Referring to FIG. 2, a preprocessing apparatus 100 for a CAD system using an active contour method according to the present invention includes an artifact removing unit 200, an outline extracting unit 300, and a suture removing unit 400.

The artificial material removing unit 200 obtains a binary image using the threshold for the input image, divides the binary image into a preset number, calculates a pixel sum with respect to the divided area, And the unnecessary artifacts are retrieved and removed from the input image using the starting point.

를 이용하여 스레쉬홀드를 적용한다. 그리고나서 도 3의 (c)에 도시된 바와 같이 이미지를 수직으로 절반으로 나누어 2개의 그룹을 얻는다. 그 후에 각 그룹에 대해 픽셀 합(pixel sum)을 계산하고, 포인터의 시작점(starting point)을 결정하기 위해 어떤 것이 더 큰 값을 가지는지 비교한다. That is, in the first step, the label and the scanned artifact are removed using thresholding. First, a mammogram image is fetched, and a threshold value is obtained to obtain a binary image as shown in FIG. 3 (b)

To apply the threshold. Then, two groups are obtained by dividing the image vertically by half as shown in FIG. 3 (c). Then calculate the pixel sum for each group and compare what has a larger value to determine the starting point of the pointer.

Through the following equation (1), the pointer can be located in the left corner of the mammogram, or in the right corner.

은 행(rows)의 최대치 수,

는 열(columns)의 최대치 수, 그리고,

는 이치영상을 나타낸다. here,

Is the maximum number of rows,

Is the maximum number of columns,

Represents a binary image.

After finding the starting point, as shown in FIG. 3 (e), unnecessary data (for example, a label or a scanned artifact, etc.) in the image is removed by moving around all the rows of the image. Therefore, after removing unnecessary data, the original mammogram image is greatly increased to obtain a desired result as shown in FIG. 3 (f).

The contour extracting unit 300 acquires an outline including the ribbone boundary information to define an energy function based on the new region, and gradually moves to the inside based on the outermost line of the input image, Upon encountering the points, the movement is stopped and the stopped points are connected to create one or more contour lines.

로 하고, 폐곡선(closed curve)을

로 하면, 다음 수학식 2와 같은 CV 모델을 이용한 새로운 영역 기반 에너지 함수를 정의할 수 있다. In other words, as a second step, the active contour method is used to obtain a contour that includes the information of the pectoral muscle boundary. Chan & Vese (CV model) proposed an active contour method based on the Mumford-Shah model. Input image

, And the closed curve

, It is possible to define a new region-based energy function using the CV model as shown in Equation (2).

,

,

, 그리고,

은 고정된 변수들이다.

은 Heaviside 함수이고,

는 새로 소개된 마스크 텀(mask term)이며, 이는 후술할 수학식 3 및 수학식 4에서 윤곽선이 내부로 이동하는 데 도움을 준다.

는 윤곽선의 정지점(stopping point)을 결정하기 위해 2개의 연속되는 윤곽선 사이의 차이를 계산하는 데 사용된다. 이는 CV 2상(two-phase) 능동적 윤곽선 모델을 다상(multiphase) 능동적 윤곽선 모델로 전환한다. here,

,

,

, And,

Are fixed variables.

Is a Heaviside function,

Is a newly introduced mask term, which helps the outline to move inward in equations (3) and (4) to be described later.

Is used to calculate the difference between two consecutive contours to determine the stopping point of the contour. This translates the CV two-phase active contour model into a multiphase active contour model.

)와 토탈 에너지 수식(

)은 아래 수학식 5 및 수학식 6에 기재되어 있다. There is a need to regularize a level set function that penalizes deviations from a signed distance function for more accurate calculations accompanied by a level set function and its evolution. The energy function of the penalizing term (

) And total energy formula (

) Are described in the following equations (5) and (6).

과

는 윤곽선(

)의 내부(inside)와 외부(outside) 각각의 이미지 인텐서티(intensity)를 계산해낸다.

에 의해 초기화된 레벨 셋 함수

는 다음 수학식 7을 참고한다. The following equations (8) and

and

Is the outline (

And calculates the image intensities of the inside and outside of the image.

Level set function initialized by

(7).

를 나타내면서, 다음 수학식 8 내지 수학식 10에 각각 기재된 변화 공식화(variational formulation)를 얻을 수 있다. The steepest descent method is used to minimize the total energy function described in equation (6)

The variational formulation described in each of the following equations (8) to (10) can be obtained.

은 라플라시안 연산자(laplacian operator)이다. 따라서 함수

는 오일러-라그랑지 방정식

을 만족시키는 함수를 최소화시킨다. 수학식 10에 기재된 마지막 텀은 구배법(gradient descent method)를 이용하여 수학식 5를 해결하면서 획득된 규칙화 텀이다. 데이터 피팅 텀

는 커브 전개(curve evolution)에 중요한 역할을 하며,

과

는 최초 텀와 두 번째 텀 사이에서 트레이드오프(tradeoff)를 지배한다. 대부분의 경우,

,

,

는 스케일링 변수(scaling parameter)로 설정하고,

는 univariate Dirac 함수를 의미한다.

함수를 최소화하는 고전적인 반복 과정은 인공 시간(artificial time)

를 이용한 다음 수학식 11의 그라디언트 흐름(gradient flow)일 수 있다. here,

Is a laplacian operator. Therefore,

Euler-Lagrange equation

Is minimized. The end term defined in Equation (10) is a rule term obtained by solving Equation (5) using a gradient descent method. Data fittings Term

Plays an important role in curve evolution,

and

Dominates the tradeoff between the first term and the second term. In most cases,

,

,

Is set as a scaling parameter,

Means the univariate Dirac function.

The classic iterative process of minimizing the function is artificial time,

And can be a gradient flow of Equation (11).

These algorithms use a second domain segmentation approach to partition objects. This divides the image into two sub-regions and continues to divide the two sub-regions into two sub-regions until an empty stopping condition is met. When the number of repetitions for obtaining the final result is set to n, n + 1 represents the total number of divided areas. FIG. 4 shows a process of multiphase segmentation using an active contour method. The contour map shown in Fig. 4 (a) shows the final contour line from the base contour line. The inner region obtained from the contour map generated through the division process is represented as a colored node generated by a solid line as shown in FIG. 4 (b). The relationship between the inner and outer regions can be expressed as Equation (12).

The contour extractor 300 according to the present invention discards the outer region and uses the inner contour to divide the contour into two portions in the next step. The similarity of each pixel within two consecutive contours is checked as follows to stop the contour at any point using the stop algorithm.

이고,

는 최종 계산된 윤곽선의 마스크 텀(mask term)이고,

는 현재 계산된 윤곽선의 마스크 텀이고,

은 원본 유방조영상 이미지 내 행(row)의 최대 수를 나타내고, 그리고,

는 열(column)의 최대 수를 나타낸다. 도 5는 계산된 윤곽선을 나타내고, 이는 흉근 경계를 포함하고 있다. here,

ego,

Is the mask term of the last calculated contour,

Is the mask term of the currently calculated contour,

Represents the maximum number of rows in the original mammogram image,

Represents the maximum number of columns. Figure 5 shows the computed contours, which include the bust breech border.

The sternal descender 400 removes the pectoral muscle by extracting a binary image of the pectoral muscle using the parameters for the mask term of the contour.

값을 이용하여 흉근을 제거한다.

는 마지막에 계산된 윤곽선의 이치영상(binary image)을 제공한다. 상술한 이치 영상을 이용하여, 흉근 이치 영상(binary image of the pectoral muscle)을 추출한다. 먼저, 도 3의 (c)에 사용된 동일한 룰(rule)을 이용하여 포인터를 이동하여 시작점(starting point)을 찾는다. 시작점을 찾은 후, 흉근은 항상 시작점과 연결되어 있다는 정보를 고려하면서 객체가 시작점과 연결되도록 하여 남아있는 영역을 제거한다. 도 6은 유방조영상으로부터 흉근이 제거되는 과정을 도시하고 있다. That is, as a third step,

Values are used to remove the pectoral muscle.

Provides a binary image of the last calculated contour. A binary image of the pectoral muscle is extracted using the above-mentioned binary image. First, a pointer is moved by using the same rule used in (c) of FIG. 3 to find a starting point. After finding the starting point, we take the information that the pectoral muscle is always connected to the starting point, and connect the object to the starting point to remove the remaining area. FIG. 6 shows a process in which the pectoral muscle is removed from the mammogram.

7 is a flowchart illustrating a preprocessing method for a CAD system using the active contour method according to the present invention. Referring to FIG. 7, when a mammogram image is input (S100), a preprocessing process (S200) is performed to increase the efficiency of detecting a region of interest (ROI) (S300) the user wants to detect. Specifically, a binary image is first obtained using a threshold for the input mammography image, a pixel value is calculated for each of the divided regions by dividing the binary image into a predetermined number, As shown in FIG. The artifact included in the inputted mammography image is removed using the determined starting point of the pointer (S210). Then, an active contour line method is used to extract contour lines including the contour line information (S220), and ultrasound contours are removed using the mask contour information of the extracted contour lines (S230).

Hereinafter, the preprocessing method for the CAD system using the active contour method according to the present invention will be compared with other existing methods, and a more accurate result image will be examined. The pre-processing method for the CAD system using the active contour method according to the present invention removes unwanted artifacts in the pre-processing step as described above.

FIG. 8 is a graph comparing a pre-processing method for a CAD system using the active contour method according to the present invention and a result image obtained by applying the other two conventional methods. The first row (row 1) represents the original mammograms obtained from the mini-MIAS database, numbered as mdb003, mdb005, mdb028, mdb044, mdb095, and mdb123, respectively. The second row (row 2) represents the ground truth of the pectoral region. The third row (row 3) and the fourth row (row 4) show images obtained by dividing the pectoral muscle by applying the two existing methods, respectively. The fifth row (row 5) shows the result of the division using the preprocessing method for the CAD system using the active contour method according to the present invention. The contour stop values used in the preprocessing method for the CAD system using the active contour method according to the present invention are mdb003, mdb028, mdb123 are 0.999, mdb005, mdb044, and mdb095 are 0.9995. All other variables were set the same.

Comparing the results of sternotomy segmentation of the mammograms mdb003, mdb005, mdb028, mdb044, mdb095, the results of the segmentation using the preprocessing method for the CAD system using the active contour method according to the present invention are compared with those of the conventional two methods And it is closer to the pesterial ground test data. In the conventional method, the pectoral muscle is linearly divided into a nearly linear shape to lose the curved pectoral region, and the Otsu threshold technique and the mathematical morphological processing are used to find a rough border of the pectoral muscle do. However, any existing method is not as accurate as the present invention. In the third row (row 3) of FIG. 8, it can be seen that there is no curve boundary of the sternum at the time of division. Therefore, the resultant image has a problem of accuracy because it includes false positive and true negative regions of linear straight line segmentation. In the segmentation result image of mdb003, the preprocessing method for the CAD system using the active contour method according to the present invention shown in the fifth column (row 5) is well divided, whereas in the fourth row (row 4) Is missing. It can be seen that the resultant images shown in the fourth column (row 4) and the fifth column (row 5) of the divided image of mdb005 are almost similar. Looking at the segmentation result image of mdb 028, it can be seen that a small hump on the sternum is missed in other methods, but is well divided in the present invention. The result images of the fourth row (row 4) and the fifth row (row 5) of the mdb 044 division result image are almost similar. In the mdb095 segmentation result image, the small details of the pectoral muscle boundaries are almost missed when other methods are applied, but are well divided in the present invention.

,

,

,

,

,

,

, 그리고,

을 이용한다. 도 9 내지 도 10은 본 발명에 따른 능동적 윤곽선 방법을 이용한 CAD 시스템을 위한 전처리 방법을 이용한 분할 결과 영상을 나타낸다. 상술한 바와 같이, mini-MIAS 데이터베이스로부터 획득된 이미지를 이용하여 본 발명을 적용하며, 정확성과 효율성에 영향을 끼치는 2가지 문제점이 있음을 전제한다. 본 발명의 정확성은 흉근의 인텐서티와 유방 영역의 나머지(rest of the breast region)는 대부분 유사하거나, 흉근의 인텐서티는 남아있는 유방 영역보다 더 낮은 경우 영향을 미친다. 도 9는 false positive와 false negative의 결과 영상을 보여주고 있다. 본 발명의 효율성은 윤곽선이 높은 인텐서티 영역 내에서 매우 천천히 움직임에 따라 높은 인텐서티 스무스(smooth) 영역내에서 유방 조영상에 영향을 미친다. 영상의 나머지 영역에서는, 본 발명은 도 10에 도시된 바와 같이 흉근의 정확한 분할 결과를 제공한다. A preprocessing method for a CAD system using the active contour method according to the present invention is applied to a mammogram image obtained from a mini-MIAS database. The intensities range of all images is 0 to 255, and each image has 1024 x 1024 pixels. At this time,

,

,

,

,

,

,

, And,

. 9 to 10 show divided result images using a preprocessing method for a CAD system using the active contour method according to the present invention. As described above, it is assumed that the present invention is applied to images obtained from the mini-MIAS database, and that there are two problems affecting accuracy and efficiency. The accuracy of the present invention is influenced by the intensities of the pectoral muscle and the rest of the breast region are mostly similar, or the intensities of the pectoral muscle are lower than the remaining breast regions. Figure 9 shows the result image of false positive and false negative. The efficiency of the present invention affects the mammogram in a high intensitized smooth region with very slow motion within the contour high intensities region. In the remaining region of the image, the present invention provides an accurate segmentation result of the sternum as shown in Fig.

The accuracy of the method according to the present invention shown in Tables 1 and 2 is calculated using the following Equation (14).

는 흉근 영역의 지상검증자료이고, 그리고,

는 본 발명에 의해 계산된 흉근 영역을 나타낸다. here,

Is the ground validation data of the pectoral region,

Represents the pectoral region calculated by the present invention.

Mammography Accuracy percent (%) 018 86.99 030 84.79 134 71.94 253 64.68

Mammography Accuracy percent (%) 006 97.70 009 98.25 132 97.35 249 98.07

The average probabilities of the accuracies calculated from Table 1 and Table 2 are 77.10% and 97.84%, respectively. The accuracy of the method according to the invention is influenced by the selection of the outline of the outline. In images with intensity intensities and images with intensities of the sternum close to the intensities of the remaining breast regions, the stop value is close to unity. However, images with significant intensities differences between the pectoral region and the remaining breast region are close to a stop value of 0.99.

The present invention can also be embodied as computer-readable codes on a computer-readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like, and a carrier wave (transmission via the Internet). In addition, the computer-readable recording medium may be distributed to a computer system connected to a wired / wireless communication network, and a computer-readable code may be stored and executed in a distributed manner.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is clearly understood that the same is by way of illustration and example only and is not to be taken by way of limitation in the embodiment in which said invention is directed. It will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the appended claims.

100: Preprocessing device for CAD system using active contour method
200: Artifact removal
300: contour extractor
400:

Claims (13)

Determining a starting point of the pointer by calculating a pixel sum with respect to the divided area by dividing the binary image into a predetermined number, obtaining a binary image using a threshold for the input image, An artifact removal unit for searching for and removing artifacts from the input image using a starting point;
And an energy function is defined by acquiring an outline including the rib boundary information, and gradually moves to the inside based on the outermost line of the input image, stops moving when the point corresponding to the predetermined reference is encountered, A contour extractor for concatenating the contour lines to generate at least one contour; And
And removing the pectoral muscle by extracting a binary image of the pectoral muscle using a parameter for the mask term of the contour. A preprocessing device for medical imaging CAD (computer-aided diagnosis) system. The method according to claim 1,
Wherein the artifact removal unit computes a pixel sum for each of the divided areas of the input image and determines a starting point of the pointer in the area where the pixel sum is largest. The computer-aided diagnosis (CAD) Preprocessing device for system. 3. The method of claim 2,
Wherein the starting point of the pointer sequentially retrieves the artifacts for all the rows of the input image. A preprocessing apparatus for a computer-aided diagnosis (CAD) system using the active contour method. The method according to claim 1,
Wherein the contour extracting unit determines a stop point of the contour through a successive contour line pixel similarity. The preprocessing apparatus for a computer-aided diagnosis (CAD) system using the active contour method. The method according to claim 1,
Wherein the stomach muscle removing unit calculates a pixel sum of the divided regions by dividing the stereoscopic two-dimensional image into a predetermined number, determines a starting point of the pointer, and removes the stomach muscles using the starting point. A preprocessing system for a computer-aided diagnosis (CAD) system. The method according to claim 1,
Wherein the input image is a mammogram image. A preprocessing apparatus for a computer-aided diagnosis (CAD) system using an active contour method. A preprocessing method performed by a preprocessing apparatus for a medical-image CAD system using an active contour method,
(a) obtaining a binary image using a threshold for the input image, dividing the binary image into a predetermined number, calculating a pixel sum for the divided area, and calculating a starting point of the pointer An artifact removal step of searching for and removing artifacts from the input image using the starting point;
(b) defining an energy function by acquiring an outline including the ribbed boundary information, moving gradually toward the inside based on the outermost line of the input image, stopping the movement when a point corresponding to a predetermined reference is encountered, An outline extraction step of connecting the stopped points to generate at least one outline; And
(c) extracting a binary image of the pectoral muscle using a parameter for the mask term of the contour line to remove the pectoral muscle. A pre-processing method for a computer-aided diagnosis (CAD) system using a method. 8. The method of claim 7,
Wherein the step (a) comprises: calculating a pixel sum for each of the divided areas of the input image, and determining a starting point of the pointer in the area having the largest pixel sum; (Preprocessing method for. 9. The method of claim 8,
Wherein the starting point of the pointer sequentially retrieves the artifacts for all the rows of the input image. A preprocessing method for a computer-aided diagnosis (CAD) system using an active contour method. 8. The method of claim 7,
Wherein the step (b) comprises determining a stop point of the contour through successive contour line pixel similarities. A method of pre-processing for a computer-aided diagnosis (CAD) system using an active contour method. 8. The method of claim 7,
Wherein the step (c) comprises the steps of: calculating a sum of pixels for the divided regions by dividing the stereoscopic binary image into a predetermined number, determining a starting point of the pointer, and removing the sick muscles using the starting point; A pre-processing method for a computer-aided diagnosis (CAD) system using a method. 8. The method of claim 7,
Characterized in that the input image is a mammogram image. A preprocessing method for a computer-aided diagnosis (CAD) system using an active contour method. A computer-readable recording medium having recorded thereon a program for causing a computer to execute a preprocessing method for a medical image CAD (computer-aided diagnosis) system using the active contour method according to any one of claims 7 to 12.

Images