JP5380185B2 - Image processing device - Google Patents

Description

  The present invention relates to an image processing apparatus for supporting image diagnosis of a lesion area.

  For example, as disclosed in Patent Document 1, there is an image processing apparatus that supports image diagnosis of a lesion area. The image processing apparatus performs image processing specified by the user on the morphological image and functional image specified by the user, thereby generating and displaying an image in which the lesion area is emphasized. Currently, various image processing functions are independently implemented in clinical applications for image processing. Therefore, the user must specify a combination of a morphological image and a functional image for extracting a lesion area from among various types of morphological images and functional images by trial and error. In addition, since there are many types of image processing for extracting the lesion area, this is also designated by the user on a trial and error basis. Therefore, the user has to select an image type or an image processing type many times before generating an intended image, which is very troublesome.

Japanese Patent Laid-Open No. 2003-126076

  The objective of this invention is providing the image processing apparatus which implement | achieves reduction of a user's effort in the image diagnosis of a lesion area | region.

  An image processing apparatus according to a first aspect of the present invention includes an image storage unit that stores data of a plurality of morphological images and data of a plurality of functional images related to a plurality of image types, the plurality of image types, and a plurality of disease names. A first table storage unit that stores a first table that associates the first table, and a first image type and a second image type that are associated on the first table with a disease name input by a user among the plurality of disease names Are read out from the image storage unit, and the image processing is performed on the first specifying unit that specifies the image type from a plurality of image types, the morphological image related to the specified first image type, and the functional image related to the second image type. An image processing unit that extracts a lesion area from a functional image related to the second image type; and a display unit that displays the extracted lesion area and a morphological image related to the first image type in an overlapping manner.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to provide the image processing apparatus which implement | achieves reduction of a user's effort in the image diagnosis of a lesion area | region.

1 is a diagram showing a configuration of an image processing apparatus according to an embodiment of the present invention. The figure which shows an example of the image type table memorize | stored in the image type table memory | storage part of FIG. The figure which shows an example of the inversion / non-inversion table memorize | stored in the inversion / non-inversion table memory | storage part of FIG. The figure which shows an example of the calculation type table between images memorize | stored in the calculation type table storage part between images of FIG. The figure which shows the typical flow of the image process performed under control of the control part of FIG. The figure which shows the flow of the image processing performed in the flow of FIG. 5 with an image. It is a figure for demonstrating the normalization process performed in step S4 of FIG. 5, and is a figure which shows the pixel value range of each image. The figure which shows the calculation procedure of the normalization scale in step S4 of FIG. The other figure which shows the calculation procedure of the normalization scale in step S4 of FIG. The other figure which shows the calculation procedure of the normalization scale in step S4 of FIG. The figure for demonstrating the calculation between images performed in FIG.5 S8. The figure which shows an example of the outline emphasis display of the lesion area | region in step S14 of FIG. The figure which shows an example of the internal highlight display of the lesion area | region in FIG.5 S14. The figure which shows the example of a GUI screen displayed by the display part of FIG.

  Hereinafter, an image processing apparatus according to an embodiment of the present invention will be described with reference to the drawings. The image processing apparatus according to the present embodiment is a computer apparatus that supports image diagnosis of a lesion area suspected of being present at a certain examination site. The examination site according to the present embodiment is not particularly limited and can be applied to any site. However, in order to specifically describe the following, it is assumed that the examination site is the brain.

  FIG. 1 is a diagram showing a configuration of an image processing apparatus according to the present invention. As shown in FIG. 1, the image processing apparatus includes an image storage unit 10, an image type table storage unit 12, an inversion / non-inversion table storage unit 14, an inter-image calculation type table storage unit 16, an input unit 18, and an image type specification unit 20. , An inversion / non-inversion specifying unit 22, an inter-image calculation type specifying unit 24, an image reading / writing unit 26, an image processing unit 28, an image composition unit 30, a display unit 32, and a control unit 34.

  The image storage unit 10 stores a plurality of morphological image data regarding a plurality of image types and a plurality of functional image data regarding a plurality of image types. Each of the morphological image and the functional image is an image generated by one modality among an X-ray diagnostic apparatus, an X-ray computed tomography apparatus, a magnetic resonance imaging apparatus, an ultrasonic diagnostic apparatus, and a nuclear medicine diagnostic apparatus. . Further, each of the morphological image and the functional image is a two-dimensional image generated by pixel value projection processing, multi-section conversion, or the like, even if it is a three-dimensional image generated by three-dimensional image processing such as volume rendering. May be. In order to specifically describe the following, it is assumed that each of the morphological image and the functional image is generated by a magnetic resonance imaging apparatus. Further, it is assumed that each of the morphological image and the functional image is a two-dimensional image.

  The image type table storage unit 12 stores a table in which a plurality of image types and a plurality of disease names are associated (hereinafter referred to as an image type table). FIG. 2 is a diagram illustrating an example of an image type table. As shown in FIG. 2, the image type table has a morphological image item and a functional image item as image items. The morphological image item is associated with an image type of a morphological image for observing brain morphological information. The functional image items are subdivided according to disease names, and image types of functional images suitable for observing a lesion region corresponding to each disease name are associated with each other. Disease names include cerebral infarction, cerebral hemorrhage, brain tumor and the like. Examples of image types include T1, T2, MRA (MR Angiography), PWI (Perfusion Weighted Image), DWI (Diffusion Weighted Image), and fMRI (functional MRI). Combinations of morphological images and functional images are ranked according to priority. The priority order depends on the use frequency of the image type for a specific disease. The higher the usage frequency, the higher the priority. For example, when cerebral infarction is to be diagnosed, the combination of the morphological image “T1” and the functional image “PWI” has the highest priority. Next is the combination of the morphological image “MRA” and the functional image “DWI”.

  The inversion / non-inversion table storage unit 14 stores a table in which a plurality of image types, a plurality of disease names, and codes indicating inversion or non-inversion of pixel values are associated (hereinafter referred to as an inversion / non-inversion table). To do. FIG. 3 is a diagram illustrating an example of the inversion / non-inversion table. As shown in FIG. 3, the inversion / non-inversion table associates a code indicating inversion or non-inversion with a combination of an image type and a disease name of a morphological image or a functional image. Inversion is one of the image processing performed to extract the lesion area corresponding to the disease name, and is to invert the pixel values of the morphological image and the functional image. Non-inversion is not to invert the pixel values. is there.

  The inter-image calculation type table storage unit 16 is a table (hereinafter referred to as an inter-image calculation type table) that associates the combination of the image type of the morphological image and the image type of the functional image with the inter-image calculation type. Remember. FIG. 4 is a diagram illustrating an example of an inter-image calculation type table. As shown in FIG. 4, the inter-image calculation type table associates the inter-image calculation type with a combination of the image type of the morphological image and the image type of the functional image. The inter-image calculation is one of image processes performed to extract a lesion area, and is a logical operation or four arithmetic operations performed on a morphological image and a functional image. Examples of the logical operation include logical sum and logical product, and examples of the four arithmetic operations include addition and subtraction. The calculation type between images is set to the calculation type between images generally used for every image type.

  The input unit 18 is an input device such as a keyboard or a mouse. The input unit 18 inputs various instructions and information from the user. For example, the input unit 18 inputs a disease name, inputs an instruction to start image processing according to the present embodiment, and specifies a position on the image.

  The image type specifying unit 20 searches the image type table using the disease name input from the input unit 18 as a keyword, and the image type of the morphological image and the image type of the functional image associated with this keyword in the image type table, respectively. Identify. More specifically, the image type identification unit 20 includes the image type of the morphological image and the image type of the functional image stored in the image storage unit 10 among the image types associated with the keyword, Identify high image types.

  The inversion / non-inversion specifying unit 22 searches the inversion / non-inversion table using the combination of the image type of the morphological image specified by the image type specifying unit 20 and the disease name input from the input unit 18 as a keyword. A code indicating inversion or non-inversion associated with the inversion / non-inversion table is specified. Similarly, the inversion / non-inversion specifying unit 22 searches the inversion / non-inversion table using the combination of the image type of the functional image specified by the image type specifying unit 20 and the disease name input from the input unit 18 as a keyword, A code indicating inversion or non-inversion associated with this keyword in the inversion / non-inversion table is specified.

  The inter-image calculation type specifying unit 24 searches the inter-image calculation type table using the combination of the image type of the morphological image specified by the image type specifying unit 20 and the image type of the functional image as a keyword. An inter-image operation type associated with the seed table is specified.

  The image read / write unit 26 reads data of the morphological image corresponding to the image type of the morphological image specified by the image type specifying unit 20 from the image storage unit 10. Further, the image read / write unit 26 reads from the image storage unit 26 data of a functional image corresponding to the image type of the functional image specified by the image type specifying unit 20. The image read / write unit 26 writes the composite image data generated by the image composition unit 30 in the image storage unit 10.

  The image processing unit 28 performs various image processing on the morphological image and the functional image read by the image reading / writing unit 26 and extracts a lesion area from the functional image. Specifically, the image processing unit 28 includes a normalization processing unit 281, an inversion processing unit 282, an alignment unit 283, an inter-image calculation unit 284, a threshold processing unit 285, and a mask processing unit 286. The normalization processing unit 281 performs normalization processing so that the pixel value ranges of the morphological image and the functional image fall within the same range. The inversion processing unit 282 inverts or non-inverts the pixel value of the morphological image according to the code specified by the inversion / non-inversion specifying unit 22. Similarly, the inversion processing unit 282 inverts or noninverts the pixel value of the functional image in accordance with the code specified by the inversion / non-inversion specifying unit 22. The alignment unit 283 aligns the same anatomical positions of the morphological image and the functional image. The inter-image calculation unit 284 performs an inter-image calculation corresponding to the inter-image calculation type specified by the inter-image calculation type specifying unit 24 for the morphological image and the functional image, and a single output image (hereinafter referred to as a calculation image). Data). The threshold processing unit 285 determines a threshold based on a pixel value at a position specified on the calculation image via the input unit 18 by the user, performs threshold processing on the calculation image with the determined threshold, and performs binarization of the binarized image. Generate data. The mask processing unit 286 performs mask processing on the functional image using the binarized image generated by the threshold processing unit 285 as a mask, and extracts a lesion area from the functional image.

  The image synthesizing unit 30 synthesizes the lesion area and the morphological image extracted by the image processing unit 28, and generates synthesized image data.

  The display unit 32 displays the composite image generated by the image composition unit 30. In other words, the display unit 32 displays the morphological image and the lesion area extracted from the functional image in a superimposed manner. At this time, the display unit 32 highlights and displays the lesion area.

  The control unit 34 functions as the center of the image processing apparatus. The control unit 34 develops an image processing program stored in a hard disk (not shown) on the memory, and controls each unit according to the image processing program, thereby executing image processing unique to the present embodiment.

  Hereinafter, clinical application examples of image processing according to the present embodiment, which are executed under the control of the control unit 34, will be described. FIG. 5 is a diagram illustrating a typical flow of image processing executed under the control of the control unit 34. FIG. 6 is a diagram showing a flow of image processing performed in the flow of FIG. 5 together with an image.

  When the user gives an instruction to start image processing via the input unit 18, the control unit 34 starts image processing according to the present embodiment. First, the control unit 34 waits for a disease name candidate to be input via the input unit 18 (step S1). The disease name may be directly input via the input unit 18 or may be selected on the disease name candidate list displayed on the display unit 32.

  When a disease name is input by the user via the input unit 18 (step S1: YES), the control unit 34 causes the image type specifying unit 20 to perform an image type specifying process (step S2). In step S2, the image type specifying unit 20 searches the image type table using the disease name input by the input unit 18 as a keyword, and specifies the image type associated with this keyword on the image type table (see FIG. 2). . At this time, the image type specifying unit 20 specifies a combination of image types having a high priority among combinations of image types of morphological images and functional images associated with disease names. When the morphological image and the functional image corresponding to the identified combination of image types are not stored in the image storage unit 10, the combination is a lower priority image type and both the morphological image and the functional image Is stored in the image storage unit 10.

  When step S2 is performed, the control unit 34 causes the image read / write unit 26 to perform image reading processing (step S3). In step S <b> 3, the image read / write unit 26 reads from the image storage unit 10 morphological image data and functional image data corresponding to the image type specified in step S <b> 2. The morphological image and the functional image are hierarchized in accordance with, for example, the DICOM (Digital Imaging and Communication in Medicine) standard and managed in the image storage unit 10. The read morphological image and functional image belong to the same examination or the same series. For example, in step S3, when the image type of the morphological image is a T1-weighted image and the image type of the functional image is a T2-weighted image, the image read / write unit 26 stores the data of the T1-weighted image and the T2-weighted image belonging to the same series. Read data and.

  When step S3 is performed, the control unit 34 causes the normalization processing unit 281 of the image processing unit 28 to perform normalization processing (step S4). In step S4, the normalization processing unit 281 normalizes both the morphological image and the functional image read in step S3. Normalization is performed in order to make the pixel value ranges of the morphological image and the functional image the same.

  FIG. 7 is a diagram for explaining the normalization process, and is a diagram illustrating pixel value ranges of the respective images I1, I2, and I3. The images I1, I2, and I3 are assumed to belong to the same series. As shown in FIG. 7, if the images are different, the pixel value ranges are also different. For example, the pixel value range of the image I1 is the minimum pixel value I1min to the maximum pixel value I1max. The pixel value range of the image I2 is the minimum pixel value I2min to the maximum pixel value I2max, and the pixel value range of the image I3 is the minimum pixel value I3min to the maximum pixel value I3max. Even if the inter-image calculation is directly performed on these images I1, I2, and I3, the extraction accuracy of the lesion area is poor because the pixel value ranges are different. For this purpose, normalization processing is performed.

  In the normalization process, first, as shown in FIG. 7, the target image of the inter-image calculation, that is, the pixel value range of the morphological image and the functional image is specified. Then, the minimum pixel value Smin and the maximum pixel value Smax in the pixel value range are specified. In other words, the minimum pixel value Smin and the maximum pixel value Smax among the pixel values of the images related to the same series are specified.

The normalization processing unit 281 normalizes the morphological image and the functional image based on the specified minimum pixel value Smin, maximum pixel value Smax, and the following expressions (1) and (2).
new_pixval = orig_pixval * fscalar (1)
fscale = 65535 / (2 * max (| Smin |, | Smax |)) (2)
new_pixval: pixel value after normalization
orig_pixval: pixel value before normalization
fscale: Normalized scale A procedure for calculating the normalized scale fscale will be described below with reference to FIGS. 8, 9, and 10. FIG. The left end FLT_MIN on the horizontal axis of the graph of FIG. 8, FIG. 9, and the upper graph of FIG. 10 is the minimum pixel value that can be taken in the floating point arithmetic, and the right end FLT_MAX can be taken in the floating point arithmetic. This is the maximum pixel value. Floating point arithmetic is an example of the data type of an image before normalization. First, as shown in FIG. 8, when the minimum pixel value Smin and the maximum pixel value Smax are specified, the larger one of the minimum pixel value Smin and the maximum pixel value Smax is specified. This process corresponds to max (| Smin |, | Smax |) in equation (2). In the case of FIG. 8, since the absolute value of the minimum pixel value Smin is larger than that of the maximum pixel value Smax, the minimum pixel value Smin is specified. Next, as shown in FIG. 9, the normalized range is calculated by multiplying the specified absolute value by 2. The normalization range is a range to be normalized. This processing corresponds to 2 * max (| Smin |, | Smax |) in the equation (2).

  Then, as shown in FIG. 10, a normalization scale fscale is calculated based on the calculated normalization range and the pixel value range of the data type after normalization. The left end SHORT_MIN on the horizontal axis of the lower graph of FIG. 10 is the minimum pixel value that can be taken in the short-precision floating-point operation, and the right end SHORT_MAX is the maximum pixel value that can be taken in the short-precision floating-point operation. The short-precision floating-point operation is an example of the data type of the image after normalization. The pixel value range that can be taken by the short-precision floating-point operation is a pixel value range that can be taken by the image after normalization, which is 65536. The normalization scale is calculated by dividing the pixel value range 65536 that can be taken by the normalized image by the normalization range 2 * max (| Smin |, | Smax |). This processing corresponds to 65535 / (2 * max (| Smin |, | Smax |)) in the equation (2).

  When step S4 is performed, the control unit 34 causes the inversion / non-inversion specifying unit 22 to perform inversion / non-inversion specifying processing (step S5). In step S5, the inversion / non-inversion specifying unit 22 searches the inversion / non-inversion table using the combination of the disease name input in step S1 and the image type specified in step S2 as a keyword, and inversion / non-inversion to this keyword. A code indicating inversion or non-inversion associated on the table is specified (see FIG. 3). Inversion or non-inversion is specified for each of the morphological image and the functional image. Inversion or non-inversion is closely related to the image operation type in the inter-image operation type table. The inversion or non-inversion determination method will be described in step S9.

  When step S5 is performed, the control unit 34 causes the inversion processing unit 282 of the image processing unit 28 to perform inversion processing (step S6). In step S6, the inversion processing unit 282 inverts or non-inverts the pixel values of the normalized morphological image and functional image, respectively, according to the code specified in step S5. For example, as shown in FIG. 6, it is assumed that the morphological image is a T1-weighted image IM and the functional image is a T2-weighted image IF. In each of the T1-weighted image IM and the T2-weighted image IF, a lesion region derived from a brain tumor is depicted. As is well known, water is drawn black in the T1-weighted image IM, and water is drawn white in the T2-weighted image IF. Therefore, the brain tumor containing a lot of water is drawn black in the T1-weighted image IM and white in the T2-weighted image IF. In this case, the inversion processing unit 282 inverts the low-pixel-value lesion area to the high-pixel value by inverting the pixel value of the T1-weighted image IM in order to set the low-pixel-value lesion area to the high-pixel value. In the morphological image IR in which the pixel value is inverted, the lesion area derived from the brain tumor is drawn white. On the other hand, for the T2-weighted image IF having a high pixel value lesion area, the pixel value is not inverted, that is, not inverted.

  When step S6 is performed, the control unit 34 causes the alignment unit 283 of the image processing unit 28 to perform alignment processing. (Step S7). In step S7, the alignment unit 283 aligns anatomically identical points of the inverted or non-inverted morphological image and functional image. This alignment is performed in order to perform an inter-image calculation (step S9) described later with high accuracy. Any existing method may be used as the alignment method. For example, there is a method based on a position marked on an image using a marker or the like at the time of image capturing.

  When step S7 is performed, the control unit 34 causes the inter-image calculation type specifying unit 24 to perform an inter-image calculation type specifying process (step S8). In step S8, the inter-image calculation type specifying unit 24 searches the inter-image calculation type table using the combination of the image type of the morphological image specified in step S2 and the image type of the functional image as a keyword. An inter-image operation type associated on the seed table is specified (see FIG. 4). The inter-image calculation is an operator for inputting a morphological image and a functional image and outputting a lesion region having a high pixel value.

  Hereinafter, the inter-image calculation will be described with reference to FIG. As shown in FIG. 11, a T1-weighted image IM and a T2-weighted image IF are considered as inputs. The image IM and the image IF are the same as the image IM and the image IF in FIG. That is, a lesion area related to a brain tumor is depicted in each of the image IM and the image IF, the lesion area on the image IM has a low pixel value, and the lesion area on the image IF has a high pixel value. The operator “+” is addition of the image IM and the image IF. The operator “−” is a subtraction between the image IM and the image IF. The operator “×” is the product of the image IM and the image IF. The operator “/” is a division between the image IM and the image IF. The operator “OR” is a logical sum of the image IM and the image IF. The operator “AND” is a logical product of the image IM and the image IF. As shown in FIG. 11, among these operators, the operator “−” (subtraction) can highlight the lesion area most clearly. Therefore, it can be said that the subtraction of the T1-weighted image IM and the T2-weighted image IF, that is, the operator “−” is most suitable for extracting the lesion area related to the brain tumor.

  Here, consider a case where the pixel value of the T1-weighted image is inverted. Adding the inverted T1-weighted image and the non-inverted T2-weighted image is equivalent to subtracting the non-inverted T1-weighted image and the non-inverted T2-weighted image. Therefore, it can be said that adding the inverted T1-weighted image and the non-inverted T2-weighted image, that is, the operator “+” is most suitable for extracting the lesion area related to the brain tumor.

  Thus, the inversion / non-inversion in the inversion / non-inversion table and the inter-image calculation type in the inter-image calculation type table are determined in relation to each other so that the lesion area is emphasized by the inter-image calculation. Here, “enhancement” means that the pixel value is remarkably high or low compared to other regions. The inversion / non-inversion and the operation type between images are determined by the user by trial and error, and can be arbitrarily updated via the input unit 18.

  When step S8 is performed, the control unit 34 causes the inter-image calculation unit 284 of the image processing unit 28 to perform inter-image calculation processing (step S9). In step S9, the inter-image calculation unit 284 performs the inter-image calculation of the type specified in step S8 on the morphological image and the functional image, and generates data of a single calculated image IO. As shown in FIG. 6, the arithmetic image IO includes a lesion area in which the pixel value is emphasized as compared with other normal areas. By performing the inter-image calculation in this way, the lesion region having a high pixel value becomes a higher pixel value. This makes it easy to extract a lesion region by binarization in the subsequent step S10.

  When step S9 is performed, the control unit 34 causes the threshold processing unit 285 of the image processing unit 28 to perform threshold processing (step S10). The threshold processing unit 285 performs threshold processing on the calculated image IO generated in step S9, and extracts a lesion area from the calculated image IO. Specifically, threshold processing is performed using the pixel value at the initial position designated by the user. As the initial position, the first point or region of interest designated on the calculation image IO by the user is used. For example, when the inside of a lesion area on the calculation image IO displayed on the display unit 32 is clicked or a region of interest is drawn by a user's mouse operation, the position is set as an initial position. Once the initial position is set, the threshold is set next. For example, the threshold value is set to a pixel value that separates a lesion area (high pixel value area) and other areas (low pixel value area) in the histogram of the operation image IO. Then, a pixel value “1” is assigned to a pixel having a pixel value equal to or greater than the threshold value, and a pixel value “0” is assigned to a pixel having a pixel value equal to or less than the threshold value. Further, a pixel area having a pixel value “1” that does not include the set initial position is assigned to a pixel value “0” as not being a lesion area. As a result, the lesion area on the operation image IO is assigned to the pixel value “1”, and the other pixel areas are assigned to the pixel value “0”. Thereby, data of the binarized image IB from which the lesion area is extracted is generated.

  When step S10 is performed, the control unit 34 causes the mask processing unit 286 of the image processing unit 28 to perform mask processing on the morphological image (step S11). In step S11, the mask processing unit 286 performs mask processing on the morphological image IM normalized in step S4 using the binarized image IB generated in step S10 as a mask, and removes a lesion region from the morphological image. Specifically, an area on the morphological image IM that is at the same coordinate position as the coordinate position of the lesion area on the binarized image IB is specified. Then, the mask processing unit 286 assigns the pixel value of the specified area to the pixel value “0”. Thereby, data of the morphological image IN from which the lesion area is removed is generated.

  When step S10 is performed, the control unit 34 causes the mask processing unit 286 of the image processing unit 28 to perform mask processing on the functional image IF (step S12). In step S12, the mask processing unit 286 performs mask processing on the functional image IF normalized in step S4 using the binarized image IB generated in step S10 as a mask, and extracts a lesion region from the functional image IF. . Specifically, an area on the functional image IF that is at the same coordinate position as the coordinate position of the lesion area on the binarized image IB is specified. Then, the mask processing unit 286 assigns pixel values other than the specified area to the pixel value “0”. As a result, functional image II from which regions other than the lesion region are removed, that is, functional image II data including only the lesion region is generated.

  When Steps S11 and S12 are performed, the control unit 34 causes the image composition unit 30 to perform composition processing (Step S13). In step S13, the image composition unit 30 synthesizes the morphological image IN generated in step S11 and the functional image II generated in step S12 to generate data of the composite image IG. The lesion area of the generated composite image IG is derived from the functional image, and other than the lesion area is derived from the morphological image. The data of the composite image IG generated by the image composition unit 30 is written into the image storage unit 10 by the image read / write unit 26.

  When step S13 is performed, the control unit 34 causes the display unit 32 to perform display processing (step S14). In step S14, the display unit 32 displays the composite image IG generated in step S13. In other words, the display unit 32 displays the lesion area extracted from the functional image so as to be superimposed on the morphological image. At this time, the display unit 32 highlights and displays the lesion area on the composite image IG. For example, as shown in FIG. 12, the display unit 32 highlights and displays the outline of the lesion area with color, brightness, or the like. In addition to emphasizing by color and brightness, emphasis may be achieved by blinking the outline. Further, as shown in FIG. 13, the display unit 32 may emphasize and display a color such as red inside the lesion area. By highlighting the lesion area in this way, the user can easily identify the boundary between the lesion area and the surrounding tissue area.

  When step S14 is performed, the image processing according to the present embodiment ends.

  As described above, the image processing apparatus holds an image type table in which a disease name and an image type suitable for diagnosing a lesion area corresponding to the disease name are associated with each other. Using this image type table, the image processing apparatus, when a disease name is input by the user via the input unit 18, is a morphological image of an image type that is most suitable for diagnosing a lesion area corresponding to the input disease name. And functional images can be automatically read out. Therefore, it is possible to reduce the trouble of selecting the image type by the user. Further, by storing the relationship between the disease name and the image type as a table, empirical knowledge regarding the selection of the image type can be stored. Further, the image processing apparatus holds an inversion / non-inversion table, an inter-image calculation type table, and the like. By using the inversion / non-inversion table and the inter-image calculation type table, the image processing apparatus can automatically select the inter-image calculation type that is most suitable for extracting the lesion area. Therefore, it is possible to reduce the time and effort required for the selection of the inter-image calculation type by the user, and it is possible to store empirical knowledge regarding the selection of the inter-image calculation type. With these various tables, a morphological image in which a lesion area is emphasized can be easily generated and displayed. Thus, the image processing apparatus according to the present embodiment can realize a reduction in the user's effort in the image diagnosis of the lesion area.

  There are cases where the automatic selection of the image type and the inter-image calculation type does not function well, or the user himself / herself wants to select. In preparation for such a situation, it is desirable in practice to have means for manually selecting the image type and the inter-image calculation type.

  FIG. 14 is a diagram illustrating an example of a GUI (graphic user interface) screen for manual selection of the image type and the inter-image calculation type displayed by the display unit 32. For example, the GUI screen is displayed by the display unit 32 during the image processing of FIG. The GUI screen has an examination site / disease name / image type selection area R1, an image selection area R2, an inter-image calculation type selection area R3, and a composite image display area R4.

  The examination part / disease name / image type selection area R1 is a part selection area R5 for selecting examination parts such as the head, chest, and abdomen, and a disease name selection area R6 for selecting disease names such as cerebral infarction, brain tumor, and cerebral hemorrhage. , T1, T2, DWI, PWI, and other image type selection areas R7 for selecting image types. By selecting these items via the input unit 18, an examination site, a disease name, and an image type are selected.

  In the image selection area R2, thumbnails S of images relating to the respective image types are displayed. When the thumbnail S is selected by the user via the input unit 18, morphological image data and functional image data corresponding to the selected thumbnail S are read from the image storage unit 10.

  The inter-image calculation type selection area R3 includes a selection area R8 for selecting an inter-image calculation type such as +,-, x, /, OR, and AND. The read morphological image and functional image are displayed in any one of the image display regions R9. Also, a weighting parameter or the like in the calculation between images can be set via the input unit 18 in the parameter region R10. In addition, when the Preview button B1 is pressed, the composite image IG generated based on the set inter-image calculation type and parameters is displayed in the composite image display region R4. By pressing the Save button B2, the set inter-image calculation type is set in the inter-image calculation type table.

  Further, the inter-image calculation type in the inter-image calculation type table may be changed according to the use frequency of the inter-image calculation type. For example, the inter-image calculation type table storage unit 16 records the number of selections of the inter-image calculation type for each image type, and specifies the inter-image calculation type having the largest selection number for every predetermined period. The inter-image calculation type table storage unit 16 updates the inter-image calculation type table by newly associating the specified inter-image calculation type and the image type with the inter-image calculation type table. By adopting such an update method, the image processing apparatus can automatically specify the inter-image operation type that is frequently used, and cause the morphological image and the functional image to execute the specified inter-image operation type. it can.

  By displaying such a GUI screen, it is possible to prepare for situations where the automatic selection of the image type and the inter-image calculation type does not function well, or when the user himself wants to select. Further, since the composite image can be simulated by manually selecting the inter-image calculation type, a more suitable inter-image calculation type can be found, which can be used for updating the inter-image calculation type table.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

  As mentioned above, according to this invention, provision of the image processing apparatus which implement | achieves reduction of a user's effort in the image diagnosis of a lesion area | region is realizable.

  DESCRIPTION OF SYMBOLS 10 ... Image memory | storage part, 12 ... Image kind table memory | storage part, 14 ... Inversion / non-inversion table memory | storage part, 16 ... Inter-image calculation seed | species table memory part, 18 ... Input part, 20 ... Image kind specification part, 22 ... Non-inversion specifying unit, 24 ... Inter-image calculation type specifying unit, 26 ... Image reading / writing unit, 28 ... Image processing unit, 281 ... Normalization processing unit, 282 ... Inversion processing unit, 283 ... Positioning unit, 284 ... Inter-image calculation , 285 ... Threshold processing unit, 286 ... Mask processing unit, 30 ... Image composition unit, 32 ... Display unit, 34 ... Control unit

Claims (5)

An image storage unit for storing data of a plurality of morphological images and data of a plurality of functional images related to a plurality of image types;
A first table storage unit for storing a first table in which the plurality of image types and a plurality of disease names are associated;
A first specifying unit that specifies, from among a plurality of image types, a first image type and a second image type associated with a disease name input by a user from among the plurality of disease names on the first table; ,
An image for reading out a morphological image related to the specified first image type and a functional image related to the second image type from the image storage unit and performing image processing to extract a lesion region from the functional image related to the second image type A processing unit;
A display unit for displaying the extracted lesion region and the morphological image relating to the first image type in an overlapping manner;
An image processing apparatus comprising: A second table storage unit that stores a second table that associates the plurality of disease names, the plurality of image types, and a code indicating inversion or non-inversion of pixel values;
The first code indicating the inversion or non-inversion associated with the combination of the disease name input by the user and the first image type on the second table is specified, and the disease name input by the user A second specifying unit that specifies the second code indicating the inversion or non-inversion associated with the combination with the second image type on the second table;
The image processing unit inverts or non-inverts the morphological image related to the first image type according to the specified first code, and the second image type according to the specified second code. The function image is inverted or non-inverted, and a specific inter-image calculation is performed on the inverted or non-inverted form image related to the first image type and the function image related to the second image type. Extracting the lesion area from the functional image related to the image type,
The image processing apparatus according to claim 1. A third table storage unit for storing a third table in which the plurality of image types and codes indicating a plurality of inter-image operations are associated;
Among the codes indicating the plurality of inter-image operations, the third code indicating the specific inter-image operation associated with the combination of the first image type and the second image type on the third table. A third specifying unit that is specified from
The image processing unit converts the specific inter-image operation corresponding to the specified third code into the morphological image relating to the inverted or non-inverted first image type and the functional image relating to the second image type. By extracting the lesion area from the functional image,
The image processing apparatus according to claim 2.   The image processing apparatus according to claim 1, wherein the display unit displays the lesion area with emphasis.   Each of the morphological image and the functional image is an image generated by one of an X-ray diagnostic apparatus, an X-ray computed tomography apparatus, a magnetic resonance imaging apparatus, an ultrasonic diagnostic apparatus, and a nuclear medicine diagnostic apparatus. The image processing apparatus according to claim 1.

Images