JP4494553B2 - Image processing apparatus and method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and method, and more particularly to an image processing apparatus and method for digitally processing medical images.
[0002]
[Prior art]
When making a diagnosis by computer tomography (CT), first, a tomographic image of a patient is taken by an image diagnostic apparatus called a modality operated by a laboratory technician. This image is displayed on a video display on the console of the device. An inspection engineer operates a film recorder called an imager connected to the modality to instruct an image necessary for diagnosis, and the instructed image is input to the imager and hard-copied to a large format film. This film is handed over to the doctor as a diagnostic image for diagnosis of the patient. Usually, several frames of images are laid out on a single film.
[0003]
Medical images used for medical treatment such as the above diagnostic images are obliged to be stored. Conventionally, a film in which a medical image is hard-copied has been stored, but recently, the law has changed, and so-called electronic storage in which a medical image is stored as digital data has become possible.
[0004]
[Problems to be solved by the invention]
When medical images are stored electronically, it is desirable that the images can be searched and displayed in units of one sheet. Furthermore, DICOM (Digital Imaging and Communication in Medicine), which is becoming an international standard for digital communication of medical images, also employs a mechanism for managing images one by one.
[0005]
Therefore, when electronically storing a plurality of medical images hard-copied on a single film, it is necessary to divide and store the images one by one. Therefore, it is necessary to cut out individual images from images recorded on the entire film read by the scanner (hereinafter referred to as “film images”).
[0006]
An object of the present invention is to cut out a small image accurately, easily, and in a short time from a medical image in which a plurality of small images are arranged substantially equally in the vertical and horizontal directions .
[0007]
[Means for Solving the Problems]
The present invention has the following configuration as one means for achieving the above object.
[0008]
An image processing apparatus according to the present invention includes an input unit that inputs a medical image in which a plurality of small images are arranged substantially equally in the vertical and horizontal directions, a binarizing unit that converts the medical image into a binary image, Detection means for detecting the existence range of the small image based on dispersion of pixel values, and dividing the binary image of the existence range of the small image into a region where white pixels are continuous and a region where black pixels are continuous, and the divided Determining means for determining a circumscribed rectangle of the area, determining a color of an area surrounded by the circumscribed rectangle having the longest side in the circumscribed rectangle as a color of a border area between the plurality of small images, and an existence range of the small image And cutting means for cutting out the small images individually based on the color of the border region .
[0009]
An image processing method according to the present invention is an image processing method of an apparatus for image processing of a medical image in which a plurality of small images are arranged substantially equally in the vertical and horizontal directions, and the medical image is input and the medical image is converted into a binary image. An area where white pixels are continuous and an area where black pixels are continuous in the binary image of the small image existing range based on a variance of pixel values of the binary image. divided into the calculated circumscribed rectangles of the divided regions, and determines the color of the border region between the circumscribing the longest circumscribed rectangle having sides surrounding the color region of the plurality in small rectangular images, the small The small image is cut out individually based on the existence range of the image and the color of the border region .
[0010]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an image processing apparatus according to an embodiment of the present invention will be described in detail with reference to the drawings.
[0011]
[Principle of image cropping]
FIG. 1 is a diagram for explaining a film image 1 read from a film.
[0012]
As shown in FIG. 1, a plurality of medical images are recorded on, for example, a half-cut film (about 35 cm × 43 cm) in a state of being arranged in a grid pattern. These images 2 have a side of several centimeters to several tens of centimeters, and are recorded on a single film so as to have the same size. Also, there is a background area (hereinafter referred to as “border”) having a width of several millimeters to several tens of millimeters as an area that separates images from each other. Since the background area is almost transparent or opaque on the film, the color of the background area of the film image 1 is almost black or white.
[0013]
In the following explanation, one to six images 2 are arranged in one film image 1 in the vertical and horizontal directions, that is, one film image 1 is described as having a maximum of 36 images 2. It is not limited to this. An area where the image 2 on the film image 1 is arranged is called a frame.
[0014]
In order to cut out individual images from the film image 1, it is necessary to detect the position of each frame on the film image 1. In the present embodiment, a border is detected, an area surrounded by the border is recognized as a frame, and the image 2 recorded in the frame is cut out. Of course, even if the film image 1 is divided by the detected border, the same result can be obtained, but in the following description, it is expressed that the image 2 recorded in the frame is cut out.
[0015]
In addition, an image is not always recorded in a frame surrounded by the detected border. In such a case, although details will be described later, the image of the frame is not cut out.
[0016]
[Configuration of image processing apparatus]
FIG. 2 is a block diagram showing a configuration example of the image processing apparatus 110 according to the present invention.
[0017]
In FIG. 2, a CPU 101 controls the entire image processing apparatus 10 via a bus 104 based on various programs stored in a ROM 102, a hard disk (HD) 107, and the like, and uses a RAM 103 as a work memory to medical images. Various image processes and communication processes including the cutout process are executed.
[0018]
An image scanner (or film scanner) 120 for reading a film image 1 or the like is connected to the scanner interface 105. An input device 109 such as a keyboard and a mouse is connected to a USB (Universal Serial BUS) interface 106 in order for an operator to input instructions and text information for processing to be executed by the image processing apparatus 110. The network interface 108 is for connecting the image processing apparatus 110 to a network 125 such as Ethernet. Connected to the video interface 111 is a monitor 127 such as a CRT or LCD used to display the acquired film image 1 and the cut out medical image, and display the processing conditions and processing status.
[0019]
The film image 1 and the cut image 2 obtained by the image scanner 120 are temporarily stored in the HD 107 and, if necessary, an image output device connected to, for example, the computers 123 and 124 and the imager 122 via the network 125 121 or a DICOM printer 126 having an interface corresponding to the DICOM protocol. The computers 123 and 124 constitute a system called PACS (Picture Archive and Communication System) for filing, viewing, and printing medical images using a network, an image database, an image viewer, an imager, and the like. The image processing apparatus 110 can also input image data sent from a PACS or the like via the network interface 108.
[0020]
The image processing apparatus 110 can be configured using a general-purpose apparatus such as a personal computer. Also, it can be configured as a dedicated device, and can be configured integrally with, for example, an image scanner, a video digitizer, a digital still camera, or a digital video camera.
[0021]
[Image clipping process]
FIG. 3 is a flowchart for explaining an example of an image cut-out procedure, which is a process executed by the CPU 101. In step S1, film image 1 is obtained from image scanner 120 or the like. The acquired film image 1 is stored in the HD 107 for subsequent processing.
[0022]
Frame determining range determination Step S2 determines a region where a frame exists (hereinafter referred to as “frame existing range”).
[0023]
FIG. 4 is a flowchart for explaining an example of the procedure for determining the frame existence range executed in step S2.
[0024]
In step S11, the film image 1 is duplicated, and the duplicate image is binarized to generate a binary image for determining the frame existence range. The reason for binarizing the image is to minimize the effects of film scratches and dirt. The frame existence range determination image may be developed on the RAM 103, which is a work area.
[0025]
Next, in step S12, the binary image is divided into a plurality of vertically elongated regions, and in step S13, the variance of the luminance values of the pixels included in each elongated region is obtained. In subsequent steps S14 and S15, the binary image is divided into a plurality of horizontally elongated regions, and the distribution of luminance values of pixels included in each elongated region is obtained. Note that the width of the elongated region is about several pixels to several tens of pixels, but is specifically determined by a trade-off between processing speed and processing accuracy.
[0026]
If the film image 1 acquired by the image scanner 120 or the like is divided into regions according to the characteristics, it can be roughly divided into an image region outside the film, an image region in the film, and a frame existence range. If the background color of the film image 1 is assumed to be black, the image area outside the film can be considered white, and the image area inside the film excluding the frame existence range can be considered black. FIG. 5 is a diagram showing an example of dispersion characteristics obtained when the film image 1 is divided into vertically elongated regions. As shown in FIG. 5, it can be determined that the region where the variance of the luminance values of the pixels is large is the frame existence range.
[0027]
In this way, in step S16, the frame existence range can be determined from the dispersion of the respective elongated regions obtained in steps S12 to S15. It should be noted that the frame existence range includes some elongated regions outside the highly dispersed region. This is in consideration of the case where there is a slight tilt or shift of the frame. Further, the subsequent processing may be performed on the determined frame existence range.
[0028]
● Border Position Detection Steps S3 and S4 in Fig. 3 detect the position of the vertical border (hereinafter referred to as "vertical border") and the position of the horizontal border (hereinafter referred to as "horizontal border"). To do.
[0029]
FIG. 6 is a flowchart for explaining an example of the detection procedure of the vertical border position executed in step S3.
[0030]
In step S21, the film image 1 is duplicated, and a binary image for border detection is generated by binarizing the duplicate image. Next, the border color (background color) is determined in step S22. This is to increase the detection accuracy by detecting the border position not only by dispersion of pixel values but also by considering the border color. Specifically, the binary image in the frame existence range is divided into a region where white pixels are continuous and a region where black pixels are continuous. Then, the circumscribed rectangles of the plurality of divided areas are obtained, the area surrounded by the circumscribed rectangle having the longest side among the circumscribed rectangles is determined as the border, and the color of the area is set to the border color (black). Or white). The reason why the border region is detected by the side rather than the area of the circumscribed rectangle is that the film image 1 having a narrow border is considered.
[0031]
Next, in step S23, the frame existence range is divided into vertically elongated regions, and in step S24, the luminance value variance of each elongated region is obtained. In step S25, the elongated region having a small variance is extracted as a temporary border region. In step S26, it is determined whether or not the color of the extracted temporary border area matches the border color. If they do not match, the temporary border area is discarded in step S27.
[0032]
Then, if it is determined by the determination in step S28 that the processing from steps S25 to S27 has been completed for all the elongated regions, a set of temporary border regions arranged at approximately equal intervals is extracted in step S29, and extracted in step S30. The position of the vertical border is detected from the set of temporary border areas.
[0033]
Note that the detection of the horizontal border position executed in step S4 only replaces the vertical / horizontal relationship in the detection of the vertical border position, and thus detailed description thereof is omitted.
[0034]
Here, the processing for detecting the border position from the extracted temporary border area will be described in detail assuming that n temporary border areas are detected with the center line of the temporary border area as the position.
[0035]
First, a temporary border region set that satisfies the condition of Expression (1) is extracted.
D (B ₁ -B ₂ ) = D (B ₂ -B ₃ ) =… = D (B _m-1 -B _m )… (1)
Where B ₁ , B ₂ , B ₃ , ..., B _m-1 , B _m are border positions, m ≤ n
D (B _i -B _j ) is the distance between border i and border j
Note that, for the convenience of determining the frame existence range described above, B ₁ and B _{m in the} equation (1) are usually located outside the frame existence range.
[0037]
Now, since there may be many sets satisfying the expression (1), further narrowing down is performed to detect (determine) the border position. The conditions for narrowing down are as follows.
(1) The number of borders should be as large as possible without exceeding the upper limit. For example, if the maximum number of frames is 6, the upper limit is 7.
(2) When the number of detected borders is m and the border interval is D, (m−1) D corresponding to the frame existence range is made as large as possible.
(3) When the width of the film is W, for example, (m−1) D> αW. Here, α is in the range of 0.55 to 0.75, for example, but is preferably about 0.6.
[0038]
The condition (1) is to prevent, for example, cutting out four images arranged in two into two. The above condition (2) is to cut out an image having as large a size as possible even when the number of borders is the same. The condition (3) is for preventing an image having an extremely small width with respect to the width W of the film from being detected.
[0039]
● Correction of border position When the border position is detected, the size of each frame is determined. Usually, the shape of the frame is a square or a rectangle close to a square. That is, the ratio H / W of the vertical and horizontal lengths H and W of the frame should be close to “1”. Therefore, the border position is corrected in step S5.
[0040]
FIG. 7 is a flowchart for explaining an example of the border position correction procedure executed in step S5.
[0041]
In step S41, the aspect ratio H / W of the frame is obtained. In step S42, it is determined whether or not H / W is smaller than β. If so, in step S43, two frames adjacent in the vertical direction are combined. For example, the border between two frames is discarded. In step S44, it is determined whether or not H / W is greater than 1 / β. If so, two adjacent frames in the horizontal direction are combined in step S45. For example, β is in the range of 0.6 to 0.8, but a desirable value is 0.65.
[0042]
Image Cutting and Blank Frame Detection Subsequently, image 2 is cut out in step S6 of FIG. Specifically, the image 2 in the frame existence range may be divided at the detected vertical and horizontal border positions.
[0043]
Next, as shown in an example in FIG. 8, since there are empty frames 3 in which some images do not exist, it is determined in step S7 whether or not the empty frame 3 has no image 2 recorded therein. Specifically, the variance of the luminance values of the entire image 2 that has been cut out is examined. If the variance is very small, it is determined that there is no image, and the image is discarded in step S8. Next, in step S9, it is determined whether or not the image 2 has been cut out from all the frames. If not completed, the process returns to step S6, and if completed, the image cutting process is terminated.
[0044]
As described above, according to this embodiment, it is not necessary to specify an image 2 to be cut out using a mouse, a digitizer, or the like, and an individual image 2 is accurately cut out from the film image 1 for easy and short-term electronic storage. can do. In particular, a large effect is expected when it is necessary to cut out image 2 from a large number of films having different image layouts for each film.
[0045]
【The invention's effect】
As described above, according to the present invention, a small image can be cut out accurately, easily and in a short time from a medical image in which a plurality of small images are arranged substantially equally in the vertical and horizontal directions .
[Brief description of the drawings]
FIG. 1 is a diagram for explaining a film image read from a film;
FIG. 2 is a block diagram showing a configuration example of an image processing apparatus according to the present invention;
FIG. 3 is a flowchart for explaining an example of an image clipping procedure;
FIG. 4 is a flowchart for explaining an example of a procedure for determining a frame existence range;
FIG. 5 is a diagram showing an example of dispersion characteristics obtained from a vertically elongated region obtained by dividing a film image in the horizontal direction;
FIG. 6 is a flowchart for explaining an example of a procedure for detecting a vertical border position;
FIG. 7 is a flowchart for explaining an example of a border position correction procedure;
FIG. 8 is a diagram illustrating a film image in which no image exists in some frames.

Claims (6)

Input means for inputting a medical image in which a plurality of small images are arranged substantially equally in the vertical and horizontal directions;
Binarizing means for converting the medical image into a binary image;
Detecting means for detecting an existence range of the small image based on a variance of pixel values of the binary image ;
The binary image of the existence range of the small image is divided into a region where white pixels are continuous and a region where black pixels are continuous, a circumscribed rectangle of the divided region is obtained, and a circumscribed rectangle having the longest side among the circumscribed rectangles Determining means for determining a color of an area surrounded by a rectangle as a color of a border area between the plurality of small images;
An image processing apparatus comprising: an extraction unit that individually extracts the small image based on an existence range of the small image and a color of the border region . Said detecting means, by said the binary image to detect an image area including the small image, detects the border region that separates the image area in the vertical and horizontal, detecting the existence range of the small image 2. The image processing apparatus according to claim 1, wherein Further, the aspect ratio H / W of the frame region surrounded by the border region is compared with a predetermined value β. If H / W <β, the border regions between the frame regions are combined to join the frame regions adjacent in the vertical direction. 3. The correction means for discarding a border region between the frame regions in order to discard the regions and to combine adjacent frame regions in the horizontal direction when H / W> 1 / β. The described image processing apparatus. It said cutting means, when the variance of the pixel values of the small image cut out is small, image processing apparatus according to any one of claims 1 to 3, characterized in that discarding the small image . An image processing method of an apparatus for performing image processing on a medical image in which a plurality of small images are arranged substantially equally in the vertical and horizontal directions,
Enter the medical image,
Binarizing the medical image into a binary image;
Detecting the existence range of the small image based on the variance of the pixel values of the binary image ;
The binary image of the existence range of the small image is divided into a region where white pixels are continuous and a region where black pixels are continuous, a circumscribed rectangle of the divided region is obtained, and a circumscribed rectangle having the longest side among the circumscribed rectangles Determining the color of the area enclosed by the rectangle as the color of the border area between the plurality of small images;
An image processing method, wherein the small images are cut out individually based on the existence range of the small images and the color of the border region . 6. A computer-readable recording medium in which a program for controlling the computer to realize the image processing according to claim 5 is recorded.

Images