JP2962920B2 - Image processing apparatus and system having the same - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the field of image processing in which an input image is scaled and output.

[0002]

2. Description of the Related Art As an example of an image output apparatus, a medical laser beam printer which records a multi-format multi-tone image on a film by scanning with a laser beam is known. In this apparatus, it is customary to record on a film an enlarged image obtained by increasing the number of pixels from the input original image and performing interpolation and enlargement. The interpolation algorithm at this time includes nearest neighbor interpolation (replication), primary interpolation (bi-linear interpolation), and cubic spline interpolation (cubic interpolation).
ic-spline interpolation) is common. These can perform smoother interpolation as the latter. The operator selects and sets which method is used, and performs an interpolation operation on the entire output image using the set method.

[0003]

In general, it is preferable to use a primary interpolation, and more preferably a cubic spline interpolation, for obtaining a smoother enlarged image as the enlargement ratio from the original image increases. However, for example, when enlarging a monotonous character string or a blood vessel image with a clear outline, a simple enlarged image can be obtained by using simple nearest neighbor interpolation. In other words, for an information portion mainly composed of characters and clear lines in the image, a better enlarged image can be obtained by using an interpolation method that does not blur the contour.

However, in the conventional apparatus, since the entire area of the image is enlarged by the same interpolation calculation method without considering the nature of the information in the image, it is better to sharply enlarge the image with clear lines. Even if there is a good part, the enlarged image of that part will have a blurred outline.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and provides an image processing apparatus capable of performing scaling processing of an input image by a scaling method suitable for the type and characteristics of the input image and outputting a high quality scaled image. The purpose is to provide devices and systems.

[0006]

According to one aspect of the present invention, which achieves the above object, a compression means for compressing an input image and a compression ratio of the compressed image compressed by the compression means are suitable for a compressed image. Discriminating means for discriminating a scaling method, scaling means for performing a scaling process on a compressed image by the scaling method determined by the discriminating means, and an output for outputting a scaled image scaled by the scaling means And an image processing apparatus.

According to another aspect of the present invention, a medical image and information indicating the type of the medical image input device that has obtained the medical image are acquired via a network. An image processing apparatus comprising: a scaling unit that performs a scaling process on a medical image according to a determined scaling method; and an output unit that outputs a medical image scaled by the scaling unit. .

According to a further aspect of the present invention, there are provided a plurality of medical image input devices, a network connected to the medical image input device, a medical image via the network, and a medical image obtained from the medical image. Information indicating the type of the medical image input device, scaling the medical image captured via the network by a scaling method determined based on the information captured via the network, and outputting the image. An image processing system comprising:

[0009]

【Example】

<Embodiment 1> The present invention can be widely applied in fields dealing with multi-tone images, such as printers, copiers, printers, workstations, and facsimile machines. As an example, an embodiment of a medical image output apparatus which is often used in the medical field and the like and prints a high-definition, multi-tone monochrome halftone image on a film or displays and outputs it on a display will be described.

In FIG. 1, reference numeral 8 denotes a semiconductor laser controller for modulating and driving a semiconductor laser, which is described in pulse width modulation, intensity modulation, or Japanese Patent Application No. Hei 1-233771 according to an input pixel signal. Modulation is performed by the method. Reference numeral 19 denotes a data processing unit which includes an interface for taking in original image data from a medical image input device (modality) 20 such as MRI, CT, CR, DSA, etc., and an image composed of a large number of pixel data. An image storage unit that stores data, an image processing unit that performs image processing such as an image arrangement and a scaling process, and a system control unit that controls the entire system are included. Details of the data processing unit 19 will be described later. Reference numeral 21 denotes an input setting unit such as a keyboard and an operation panel, and the operator inputs types such as an output format and an interpolation method. 22 is C
It is a display for display such as an RT display or a liquid crystal display, and displays and outputs an image subjected to interpolation and enlargement processing according to a set format. The display 22
Can also display the input original image and is used for editing a print image.

Next, the configuration of the printer unit will be described.
Reference numeral 1 denotes a semiconductor laser, and 2 denotes an optical system such as a collimator lens, which converts light from the semiconductor laser 1 into parallel light. 3
Is an aperture stop, 4 is a beam splitter, 6 is a condenser lens,
Reference numeral 7 denotes a photodiode. The output of the photodiode 7 is input to a semiconductor laser controller 8, and monitors the intensity of the laser beam split by the beam splitter 4. On the other hand, a lens 5 and a rotating polygon mirror 9 as a deflecting means for performing main scanning are provided in a direction in which the beam splitter 4 is transmitted straight.
Is arranged. 10 is an fθ lens for tilt correction, 1
1 is a folding mirror for directing the direction of the light beam in a direction perpendicular to the film 12 as a recording medium. Reference numeral 16 denotes a supply magazine for accommodating a large number of unused films, reference numeral 17 denotes a receive magazine for accommodating photosensitive-recorded films, and reference numeral 13 denotes a motor for performing sub-scanning. Reference numeral 14 denotes a roller which is connected to the motor 13 and performs sub-scanning of the sheet-like film 12. An encoder 15 is attached to a rotation shaft of the roller 14 to detect a rotation state of the roller 14. As the encoder 15, for example, a laser rotary encoder or the like is suitable. The film 12 is taken out of the supply magazine 16 by a take-out mechanism (not shown) and sent to a roller 14. The film 12 is sub-scanned at a low speed by the roller 14, and is simultaneously scanned two-dimensionally on the film 12 by a modulated light beam that is main-scanned. The latent image is recorded by exposure. The recorded film is stored in the receiving magazine 17. Although not shown, a developing device for automatically developing the recorded film is provided, and the recorded film can be selectively sent to either the receiving magazine 17 or the developing device. I can do it. Reference numeral 18 denotes a photodiode for obtaining a signal (BD signal) indicating the start of main scanning for synchronizing each main scanning. The semiconductor laser controller 8 modulates and drives the semiconductor laser 1 according to the pixel signal sent from the data processing unit 19 while synchronizing with the output of the photodiode 18. In order to obtain the timing at which each scanning line starts to be drawn based on the BD signal, it is necessary to obtain the BD signal at as accurate a timing as possible in order to draw a highly accurate image. Therefore, when the scanning light beam passes through the photodetector 18 and detects a signal, the semiconductor laser 1
Are configured to continuously oscillate at a constant output. The optical output of the semiconductor laser 1 is forcibly stopped during a blanking period other than when the light enters the photodetector 18 in order to prevent irregular reflection from occurring at the corners of the rotating polygon mirror. .

Next, the details of the data processing section 19 will be described. FIG. 2 is a block diagram mainly showing the data processing unit 19. Reference numeral 31 denotes a system control unit, which includes a CPU 32 and a system memory 33. The system control unit 31 performs overall control based on various settings input from the input setting unit 2. An image storage unit 34 includes a compression circuit 35, a compression table 36, and a memory unit 37 such as a semiconductor memory device or a magnetic memory medium. Reference numeral 38 denotes an image processing unit, which includes a decompression circuit 39, an interpolation coefficient control circuit 40, a line buffer 41, and an interpolation calculation unit 42. This interpolation operation unit 4
Semiconductor laser controller 8 according to the output value from
To drive the semiconductor laser 1 to generate a modulated light beam for recording. Alternatively, an image is displayed on the display 22. 44 is an interface for MRI, CT,
It has a receiving circuit for receiving image data sent from the medical image input device 20, such as a CR or DSA. Each unit is interconnected by a system bus 46.

The image data inputted by the image input device 20 is taken in via the interface 44 under the control of the CPU 32, and the data compressed by the compression circuit 35 is written in the memory 37 of the image storage 34. Compression circuit 35
According to a method of encoding a difference from a previous pixel value in a unit of one line in accordance with the number of pixels of an image preset by the CPU 32 (a detailed algorithm will be described later), the compressed data is stored in the memory 37. While writing to
The address on the memory 37 is stored in the compression table 36. From the upper left pixel to the lower right pixel of one image, the image is sequentially transferred. When the transfer of one image is completed, the CPU 32 determines the compression ratio (compressed data) for each line based on the compression table 36 for one line. (Data amount / original data amount) is calculated and stored in the system memory 33. When outputting a plurality of images in a multi-format on a single film or a display, the above operation is repeated to acquire a plurality of images.

The operator sets an output format (output arrangement of a plurality of images) and an interpolation processing method from the input setting section 21. In the apparatus of the present embodiment, three types of interpolation processing methods (1) nearest neighbor interpolation, (2) cubic spline interpolation, and (3) selective interpolation are prepared, and the operator selects one of these.
The interpolation methods (1) and (2) are the same as the conventional ones, and when these are selected, the same interpolation method is performed over the entire film area as in the conventional case.

The selective interpolation of (3) is a method of switching an interpolation algorithm depending on the case. Specifically, the characteristics of the image are determined for the entire image or for each part, and the characteristics of the determined image are determined. A suitable interpolation method, here, either cubic spline interpolation or nearest neighbor interpolation is selected and adopted. Details of this selection interpolation are as follows.

Various parameters can be considered as parameters representing the properties of the image (the monotony of the image). In this embodiment, attention is paid to the compression ratio. In general, when arranging characters in an image, arrange them on a uniform part of the background around the film,
For example, in the case of an 8-bit image, an extreme difference of 255 from the background 0 is often set as the density value. In the case of a differential compression algorithm for compressing a pixel with a difference from a previous pixel value in the main scanning direction, a short code is assigned to a difference of 0, ± 1, ± 2, ± 3, etc., and high compression is performed. Where -1,-
Since 2 and -3 correspond to +255, +254, and +253, respectively, high compression can be obtained by differentially compressing an area including a character. On the other hand, even for a clear line, high visibility can be obtained because the fact that the background is monotonous means that the background is monotonous. High compression (small compression ratio)
Sites can be determined that monotonic image, by comparing the threshold C _T and the compression ratio is in the selected interpolation of this embodiment, high compression (the compression ratio is small) part that is not blurred contour interpolation An enlargement method (nearest neighbor interpolation) is selected, and a smooth interpolation enlargement method (cubic spline interpolation) is selected for other low compression (high compression ratio) parts.

The threshold C _T of determining compressibility properties of the image varies depending on the assignment side of the compression codes. In this embodiment, empirically, the maximum compression rate is 12.5% (the data is 1
/ 8 is compressed) set C _T = 15% in the case of code, and, if the maximum compression ratio of codes of 25% (data is compressed to 1/4 at the maximum) C _T = It is set at 28%. The calculated compression ratio is larger when the cubic spline interpolation than this threshold C _T, otherwise selects the nearest neighbor interpolation.

In FIG. 2, when the printing operation starts, C
The PU 32 extracts the compressed image data stored in the memory 37 and transfers it to the expansion circuit 39, expands the data compressed for each line to reproduce the original data, and writes it to the line buffer 41. When the format is set so that a plurality of images are arranged and output in the main scanning direction, the data of each image in the memory 37 is extracted one line at a time, and the data obtained by arranging the data according to the format is obtained in the line buffer 41.

The magnification of the image is determined by the number of pixels of the original image,
It is determined in consideration of the film size, the set output format, the border, and the like. This is described in, for example,
No. 53682 is used.

The choice in the interpolation is determined by the CPU32 according comparison between the compression ratio and the threshold C _T of each line of the operational method each original image of the interpolation, as a result of the interpolation coefficients to the interpolation coefficient control circuit 40 Is set. Details of the interpolation coefficient control circuit 40 will be described later. In the interpolation calculation unit 42, based on the contents of the line buffer 41,
Using the interpolation method based on the interpolation coefficient set in the interpolation coefficient control circuit 40, the image is enlarged at the determined enlargement ratio. Since this processing is performed by hardware using an arithmetic circuit employing a so-called pipeline method, the processing speed is high and a calculation result is output in real time. Therefore, a large-capacity frame memory for printing or display is not required. This pipeline type arithmetic circuit can change the interpolation method only by changing the interpolation coefficient to be given. The specific hardware configuration is described in, for example,
These are disclosed in JP-A-3-49983, JP-A-63-49972, and JP-A-63-53683. The enlarged image data output from the interpolation calculation unit 42 in real time is transmitted to the semiconductor laser controller 8 and the display 22.
And the image is output by raster scan.

The interpolation processing is not limited to the method of calculating by a pipeline. A page memory is prepared, an image to be output subjected to scaling / arrangement processing is temporarily stored in the page memory, and then the contents are printed. You may do it. The page memory is constituted by a semiconductor memory having a capacity for one output image. For example, the image output matrix is 4096 ×
If 5120 pixels × 12 bits / pixel, the capacity of the page memory is about 30 Mbytes.

FIG. 3 shows a specific configuration of the interpolation coefficient control circuit 40. The BD signal 132 is an output signal from the photodiode 18 of FIG.
Is used as a signal for synchronizing data transmission. The counter 133 is cleared by the pulse of the BD signal 132, and thereafter, counts up the clock by the number of pixels for one scan. The output of the counter 133 is a CPU
The values of four registers 134 to 137 for setting a change point of the interpolation coefficient set by
38. The outputs of the comparators 138 are respectively input to the selection circuits 144, and when the respective change points are reached, four registers 140 for selecting an interpolation coefficient are used.
The interpolation coefficient selection parameters set to 143 are output from the selection circuit 144 and sent to the interpolation calculation circuit.
The initial value of the interpolation coefficient selection parameter is set in the register 139.
The value set in advance is used. In this way, the interpolation coefficient in the main scanning direction is controlled for each original image. The selection of the interpolation coefficient in the sub-scanning direction is performed for each line. In this example, the number of output images arranged in the main scanning direction is four or less. However, if the number is more than four, it can be dealt with by increasing the number of registers and comparators in FIG.

FIG. 4 shows an example of the output of the system according to the present embodiment, and shows an example of how the interpolation method is switched for each part of the image with respect to the film 12. With respect to each of the four images 51, 52, 53, and 54, 56, 59, 62, and 65 are low-compression (high compression ratio) areas in which a complicated image is drawn. Interpolation is applied and smooth interpolation is performed. Other 55, 57, 58, 59, 61, 6
Since areas 4 and 66 are high-compression (low compression ratio) areas centered on monotonous character information and the like, nearest neighbor interpolation is applied, and a sharp enlarged image is obtained in these areas. As described above, in the main scanning direction, the optimum interpolation method is selected for each original image constituting the multi-format, and in the sub-scanning direction, the optimum interpolation method is selected for each line of the original image. For simplification, an interpolation method may be selected for each original image in the sub-scanning direction.

Next, the differential compression algorithm used in the present embodiment will be described more specifically. Assuming an image (x (0) to x (255)) in which one line consists of 256 pixels,
Consider encoding (Huffman code) the difference d between the pixel value of a certain coordinate and the pixel value of the previous coordinate.

The density of each coordinate is represented by 8 bits (256 gradations), and the head pixel x (0) of one line
For, there is a previous pixel, so it is determined that the previous value is zero. d can be represented as follows.

[0026]

[Outside 1]

Taking the above d statistics for medical images handled in the embodiment, the distribution often becomes almost as shown in FIG. The distribution in FIG. 5 indicates that the difference d from the pixel value of the previous coordinate is extremely high, such as 0, ± 1, ± 2, ± 3. Therefore, for such a distribution,
If a code shorter than 8 bits is assigned to a frequently occurring value such as a difference d of 0, ± 1, ± 2, ± 3, ± 100,
Even if the code length assigned to infrequent values such as ± 200 becomes 15 bits or 16 bits, it is possible to reduce the total number of bits for the entire image. That is, data can be compressed.

The range of the difference d from the previous pixel value is -255 ≦
d ≦ 255, but the difference d ′ (0 ≦ d ′ ≦ 2
55), it can be represented by 8 bits. That is, when an arbitrary pixel value x _i is considered, the range in which the difference d _i exists is not 511 points of −255 ≦ d _i ≦ 255, but x
There are 256 points of _i- 255 ≦ d _i ≦ x _i , which can be represented by 8 bits.

[0029]

[Outside 2] This can be seen from the table below, where d = 1 and d = −
255 indicates that the processing is the same.

[0030]

[Table 1] FIG. 6 shows the frequency of d 'based on this table. If a code is created from this distribution, +255, +254, +253
Also, the code length becomes shorter as in +1, +2, +3. For this reason, an image of a character written with a density of 255 in an area with a density of 0 in the background or an image with many clear lines is highly compressed (small compression ratio).

The code length is determined by the distribution of d 'in the case of the Huffman code. Here, the compression ratio is considered when the code length of d' = 0 is 1 or 2. In the above example, when the data is all 0 (x (i) = 0, I = 0 to 255), if the code length of d ′ = 0 is 1, １ ／ = 0.125
Therefore, if the compression ratio is 12.5% and the code length of d '= 0 is 2, the compression ratio is 25% from 2/8 = 0.25. Threshold C _T determines the nature of the image, it is necessary to vary depending on the code length of the above d '= 0. In this embodiment, empirically, when the compression ratio of d '= 0 is 12.5%, C _T = 15%,
d '= 0 the compression rate when 25% is set as the C _T = 28%.

<Embodiment 2> As another embodiment, an example using DCT transform (discrete cosine transform) will be described. In the previous embodiment, compression was performed in units of one line. However, in the DCT conversion in this embodiment, an original image is divided into a plurality of blocks of N × M pixels, and DCT conversion is performed in units of blocks and compressed. For example, 256 pixels x
The original image of 256 pixels is divided into blocks of 8 pixels × 8 pixels, and compression is performed in block units. Even in this case, since a monotonous image exhibits a high compression property as in the previous embodiment, the interpolation calculation method is selected according to the compression ratio.

The block diagram of the configuration of the present embodiment is almost the same as that of FIG. 2 of the previous embodiment, and different portions will be mainly described with reference to FIG. In this embodiment, the member 35 shown in FIG.
A compression circuit, 36 is an address table, and 39 is a DCT decompression circuit. The image data is sequentially written in the DCT compression circuit 35 from the upper left to the lower right of the image. The DCT compression circuit 35 converts the input original image into 8 pixels ×
It is divided into 8 pixel blocks, and DCT is applied to each block.
Perform compression and connect it in the line direction to 256 pixels x
As an image data of eight lines, an address table 36 for managing addresses on the memory 37 in a data unit of 256 pixels × 8 lines is created. Then, the compression ratio is calculated for each data unit. 256 pixels x 256 pixels (line)
In this case, the address table 36 has 32 addresses.

The DCT compression is performed according to the following procedure. Note that this DCT compression is equivalent to the method defined by JPEG. (1) A constant value (128 or the like) is subtracted as a DC component from each pixel of 8 × 8 pixels in one block. (2) Perform two-dimensional DCT transform of 8 × 8 pixels. (3) The above-mentioned D is obtained by a quantization coefficient determined empirically.
The 8 × 8 pixel points subjected to the CT conversion are divided and quantized. Here, since the image deterioration is remarkable depending on how to select the quantization coefficient, a quantization coefficient with little image deterioration is selected. (4) The quantized 8 × 8 pixel is scanned in an oblique direction from the upper left to the lower right in a zigzag manner, and the length of the continuous 0 and the value of the pixel where the continuous 0 ends are paired and encoded by Huffman coding. And compress the data.

The calculation of the compression ratio representing the characteristics of the image is performed for each block, and the comparison between the compression ratio and the threshold value is performed on a line basis (for eight lines). That is, the average value of the compression ratios of the eight horizontal blocks is obtained, and the magnitude of this is compared with the threshold value to determine the interpolation calculation method. The interpolation method is not limited to the method of determining the interpolation method for each line, and the interpolation method may be switched for each block by comparing the threshold value for each block.

When outputting an image for printing or display, the data is read from the data memory 37 in units of eight lines based on the contents of the address table 36 and written into the DCT decompression circuit 39. The DCT decompression circuit 39 processes and decompresses the data in a procedure completely opposite to the previous compression, and reproduces the original data. Based on this, an interpolation operation is performed by a pipeline method in the same manner as in the previous embodiment, and the result is output.

Various modifications and developments are conceivable for the above embodiments. For example, primary interpolation may be used instead of cubic spline interpolation as a method for performing smooth interpolation.

When the interpolation method is determined, the interpolation method may be determined only when N (N ≧ 2) or more identical compression areas continue in the original image.

[0039] Further, the threshold C _T in the above embodiment has been set to a constant value, may be changing the value of C _T in accordance with the enlargement ratio threshold C _T as a function of the image magnification. Furthermore, it is also possible to vary the C _T using a parameter other than magnification.

In the above embodiment, two kinds of interpolation methods are switched in each part of the image. In a part where the interpolation method changes, an intermediate interpolation coefficient between the nearest neighbor interpolation and the cubic spline interpolation is given. If an intermediate interpolation method is adopted, an extremely high quality enlarged image that changes very smoothly can be obtained.

Further, image scaling is not limited to enlargement. For example, pixel interpolation is necessary even in the case of a non-integral reduction,
In this case, the interpolation method may be switched according to the properties of the image.

In the above description, the compression ratio is used to determine the properties of the image. However, the properties of the image may be automatically determined using other parameters. Further, the operator may determine the nature of the image by himself, and instruct the system using a pointing device or the like.

In each of the above embodiments, the processing of a monochrome halftone image has been described. However, when the image to be handled is a color image, it can be achieved by expanding the apparatus of the above embodiment. Generally, color images are RG
The above-described image processing is performed for each of the three systems of the R image, the G image, and the B image. If the final output image is a binary image, a method such as dither processing may be added.

<Embodiment 3> Next, an embodiment of an image processing network system having an image output device as shown in the above embodiment will be described. This is an image processing network that can be used by many people to edit or print out a plurality of images obtained by various medical image input devices (modalities), and is highly useful in facilities such as hospitals that are involved in the medical field.

Images obtained by various types of medical image input devices have different properties, and it is preferable to switch the image scaling method according to the type of the input device. For example, X-ray C
Since an image obtained by a T-scanner or MRI is a dense image with a large amount of information of multiple gradations, smooth interpolation by cubic spline interpolation is suitable. On the other hand, ultrasonic diagnostic equipment and D
Since the image obtained by the SA device is coarse and has a small number of gradations, sharp interpolation by nearest neighbor interpolation is suitable. Further, since an image obtained by a DR (Digital Radiography) device is an image having the above-mentioned intermediate properties, primary interpolation is suitable. Therefore, in this embodiment, an original image is scaled by automatically setting a scaling method suitable for the type of the image input device used.

FIG. 7 is an overall view of the network of this embodiment. In the figure, 101 is an optical fiber loop for integrating the entire network, 102 is a loop interface, 10
Reference numeral 3 denotes a LAN, and 104 denotes a high-speed digital line, which is connected to another network. Reference numeral 105 denotes a network interface, 106 denotes an image management computer, and 107 denotes a large-capacity secondary storage device such as an optical disk. 108, 1
Reference numerals 09, 110, 112, and 113 denote medical image input devices (modalities) for generating digital images.
A line CT scanner, an MRI apparatus, a DSA apparatus, a DR (digital radiography) apparatus, and a CR (computed radiography) apparatus. Images obtained by these image input devices are stored in the storage device 107 via the image management computer 106. Reference numeral 111 denotes an image output device, specifically, a laser printer for recording an image on a film in a multi-format as described in the above embodiment. Reference numeral 114 denotes an image reading device, which is a film digitizer that scans and scans a photographed X-ray film or the like and converts it into digital image data. 11
Reference numeral 5 denotes an image workstation for an operator to issue an image processing command. By preparing a large number of such workstations, a large number of operators can use the network simultaneously.

Next, the configuration of the image output device 111 will be described with reference to FIG. Reference numeral 116 denotes a data processing unit of the image output device. The data processing unit 116 includes a system control unit 118,
It comprises an image storage unit 119, an image processing unit 120, a network interface 121, and a system bus 117 connecting these four.

The system control unit 118 comprises a CPU 127 and a system memory 128, and receives a command from the network interface 121 and the input setting unit 123.
It stores and reads image data into and from the image storage unit 119, sets interpolation coefficients and their selection parameters to the image processing unit 120, and issues an operation command to the printer mechanism control unit 122. The image processing unit 120 includes an interpolation coefficient control circuit 129, an interpolation processing circuit 130, and an interpolation line buffer 131. The data output from the interpolation processing circuit 130 is displayed on the display 124 and the LD controller 12.
5 is performed. The network interface 121 has a function of receiving a packet transferred from the network 126 and transferring the packet to the system control unit 118 and a function of writing image data to the image storage unit 119.

Here, a method of selecting an appropriate scaling method for each original image will be described. In this embodiment, an interpolation method is selected based on information transferred to the network. Information transferred to the network is performed in units called packets. The data structure of this packet includes a packet identifier 146 and a data part 147 as shown in FIG. In this embodiment, there are six types of identifiers as shown in FIG.

When outputting an image, first, a format specification packet is transmitted from the image workstation 115 to the network, and the image output device 111 receives the packet. The format here means the number of frames M in the horizontal direction when a plurality of images are output in a multi-format.
And the number of frames N in the vertical direction.

Subsequently, various packets are sequentially transmitted from the image workstation 115 for each image (M × N) constituting the multi-format image. These are a position designation packet, a modality designation packet, a size designation packet, an interpolation method designation packet, and an image data packet. The position specification packet specifies a position in the M × N format where an image is to be placed. The modality designation packet designates a modality that has obtained an image from the types shown in FIG. 11 connected to the network. In the size specification packet, the horizontal direction of the transferred image,
The number of each pixel in the vertical direction is specified. In the interpolation method designation packet, designation is performed from among the types of interpolation methods as shown in FIG. In this case, the interpolation method is determined according to the type of the modality based on the correspondence table as shown in FIG. The CPU of the image output device calculates a scaling factor determined from the format specified by the packet and an interpolation coefficient to be given to the arithmetic circuit determined by the specified interpolation method before image output is started.

As for the image data packet, the data of the original image is generally 256 × 256 pixels to 1024 pixels.
Since it has a large capacity of × 1024 pixels, image data for one line is repeatedly transferred as one image data packet by the number of lines of the image. The image packet is transferred from the image management computer 106 or from each modality to the image output device 111 by a remote command from the image workstation 115.

The image output device 111 outputs the original image M ×
After receiving the above N packets from the network, the printing operation is started in response to an image output start command issued by the operator.

<Fourth Embodiment> In the third embodiment, an interpolation system designating packet is transferred to a network as information for determining a scaling system for each image, and the interpolation system is determined by this. As an alternative, a packet indicating the correspondence between the modality and the interpolation method as shown in FIG. 13 may be transmitted before the start of image output, without using the interpolation method designation packet. In this case, the CPU of the image output device
Uses the received correspondence table to determine an interpolation method suitable for the modality specification packet transferred for each original image and calculate an interpolation coefficient.

As a further alternative, the contents shown in FIG. 13 are not transferred as packets but are stored in advance in the image output device 111 as a table, and a table based on the contents of the modality designation packet transferred for each image is stored. May be used to determine the interpolation method for each image. That is, the modality specifying packet is information for determining the scaling method.

<Embodiment 5> Packets include text (character) packets in addition to the types shown in FIG. Basically, for a text which is a binary level image, sharp scaling by nearest neighbor interpolation is appropriate. The display of such a text uses a technique of overlaying and displaying a multi-valued image obtained by the modality, a so-called overlay technique. If the overlay is to be executed in the above embodiment, the overlay processing is performed directly on the image data in the memory 19, so if the image is subjected to cubic spline interpolation, the characters will also be cubic splined. Therefore, a configuration for avoiding this is shown in FIG.

In FIG. 14, the image processing unit 158 comprises two systems of image data interpolation units 159, 160, 161 and overlay data interpolation units 162, 163, 164. The two interpolations are pipelined with the same clock, and the OR circuit 165 performs an OR for each bit. This makes it possible to combine the image data and the overlay data by performing different interpolation processes. This output is sent to the display 124 or the LD controller 125, and an image is output.

[0058]

As described above, according to the present invention, a high-quality scaled image can be obtained by switching the scaling method according to the nature and type of the image.

[Brief description of the drawings]

FIG. 1 is an overall system configuration diagram of an embodiment of a medical image output apparatus.

FIG. 2 is a main block diagram of the apparatus of FIG. 1;

FIG. 3 is a block diagram of a circuit that performs switching control of interpolation coefficients.

FIG. 4 is an explanatory diagram for switching an interpolation method depending on a region on a film.

FIG. 5 is a diagram illustrating an example of a frequency of a density difference between adjacent pixels.

FIG. 6 is a diagram illustrating an example of a frequency of a density difference between adjacent pixels.

FIG. 7 is an overall configuration diagram of an embodiment of a medical image processing network system.

FIG. 8 is a configuration diagram of an image output device.

FIG. 9 is a diagram showing a data structure of a packet.

FIG. 10 is a table showing types of packet identifiers;

FIG. 11 is a table showing types of modalities.

FIG. 12 is a table showing types of interpolation methods.

FIG. 13 is a correspondence table between modalities and interpolation methods.

FIG. 14 is a diagram illustrating a modification of the configuration of the image processing unit.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Semiconductor laser 3 Aperture stop 4 Beam splitter 7 Photodiode 8 Semiconductor laser controller 9 Rotating polygon mirror 10 fθ lens 13 Motor 14 Roller 15 Encoder 16 Supply magazine 17 Receive magazine 18 Photodiode 19 Image processing unit 20 Image input device 21 Input setting unit 22 display 31 system control unit 32 CPU 33 system memory 34 image storage unit 35 compression circuit 36 compression table 37 memory 38 image processing unit 39 expansion circuit 40 interpolation coefficient control circuit 41 line buffer 42 interpolation calculation unit 44 interface 46 system bus

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-61-201591 (JP, A) JP-A-2-9268 (JP, A) JP-A-2-67077 (JP, A) JP-A-2- 201582 (JP, A) JP-A-4-96876 (JP, A) JP-A-5-12352 (JP, A) JP-T4-500421 (JP, A) International publication 90/16034 (WO, A2) ( 58) Fields surveyed (Int.Cl. ⁶ , DB name) G06T 3/40 A61B 5/055 A61B 6/03 G09G 5/36 H04N 1/393

Claims (8)

(57) [Claims] A compression means for compressing an input image;
Determining means for determining a scaling method suitable for the compressed image based on the compression ratio of the compressed image compressed by the reducing means;
An image processing apparatus comprising: a scaling unit that performs a scaling process on a compressed image by a different scaling method; and an output unit that outputs a scaled image that has been scaled by the scaling unit. 2. The apparatus according to claim 1 , further comprising storage means for storing a compressed image.
The magnification / reduction means corresponds to the compressed image fetched from the storage means.
The image processing apparatus according to claim 1, wherein the image processing device performs a scaling process . 3. The method according to claim 1, wherein each of the portions of the compressed image includes
3. The image processing apparatus according to claim 2, wherein the determination is made and the scaling is performed by said scaling means . 4. The method according to claim 1, wherein the input image is a medical image.
3. The image processing apparatus according to any one of claims 1 to 3, 5. The medical image and the medical image obtained therefrom.
Network information that indicates the type of medical image input device
And Tokomu capturing means through a scaling means for scaling the medical image by the scaling method is determined based on the information, and outputs the medical image scaling in the scaling means output And an image processing apparatus. Wherein said capture means preparative plurality of medical images
Inclusive, said output means is an image processing apparatus according to claim 5 for outputting a format which is set a plurality of medical images. 7. The image processing apparatus according to claim 6 , wherein a scaling ratio of the scaling unit is determined based on the set format. 8. A medical image input device, comprising : a plurality of medical image input devices;
A network connected to the image input device for
Medical image and the doctor who obtained the medical image through the network
Information indicating the type of the medical image input device is fetched, and the network
Determined based on the information captured through the network
Captured via the network by the variable magnification method
Have the image output device to output as the row variable magnification of medical image
An image processing system, comprising: