JPH0310381A - Device and method of displaying picture - Google Patents

Abstract

PURPOSE: To make a user vary the window and the level of a picture by replacing the pixel groups, which define one picture, with the values corresponding to the pixel groups led from other transformation functions and redisplaying the picture in accordance with the transformation functions which are defined by the same one to one function. CONSTITUTION: When the user moves a screen cursor to another picture and starts to change the window and the level of the picture, the values of the pixels, which define the picture under the cursor, are replaced by the values defined by the corresponding values contained in a reference table (LUT)-1. Then, the picture is assigned to a LUT-2 and the LUT uses the eight bit pixels, which currently define the picture under the cursor, so as to map the pixels to itself. In other words, the LUT-2 defines the same 1:1 function. Then, the LUT-1 is assigned to the picture under the cursor. Thus, the user changes the window and the level of the picture.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing system. In particular, it relates to a system that allows a user to change the window or level of an image displayed on a display device.

[Prior Art] In a picture recording and communication system (PAC5), a so-called medical image obtained from a diagnosis, such as a computerized tomography scanner, consists of a series of picture elements (pixels) representing each position within the image. Represented by

The actual value of the pixel is quantized using, for example, 12 bits to represent the individual intensity levels. This value is generally called the Blaiskel value. Therefore, a 12-bit pixel has 40 pixels that vary in brightness from black to white.
Represents one of 96 grayscale values. However, many display devices use an 8-bit gray scale. Therefore, each 12-bit pixel must be converted to an 8-bit value before the pixel is displayed on the display.

To this end, many image processing systems use so-called look-up tables (LUTs) to convert 2-bit pixels into 8-bit pixels. Here, L
The UT is defined according to the individual window and level values. Typically, such a system has four LUs.
T is used and several bits are added to each 12-bit pixel, for example 2 bits I, to select the address of one LUT. The fact that the system uses four LUTs generally means that the user can independently change the windows or levels of the four images being displayed. Therefore, a system with only four columns that can change the window or level of the displayed image is limited. This limitation becomes significant in the example where many images, for example 20 images, are displayed on the display and the user changes the windows or levels of more than four images. In such an example, at least 5 LUT address bits and 20 LUTs are required to provide independent window and level control for each displayed image.

However, adding three LUT address bits requires an expensive redesign of some aspects of the PAC3.

Adding 16 LUTs is also quite expensive.

SUMMARY OF THE INVENTION The above limitations are limited to only two LUT address bits and four LUT address bits.
by using only three of the LUTs. In particular, LUT-0 is used as a default LUT and is assigned to each image as it is displayed on the display. Subsequently, when the user moves the screen cursor over one of the images and begins changing the window or level of that image, that image is reassigned from LUT-0 to LUT-1.

LUT-1 then converts the 2-bit pixel of that image into a corresponding 8-bit value, depending on the current window value set for that image, as done in the previous configuration. used to reconvert to . Subsequently, when the user moves the screen cursor to another image and begins changing the window or level of that image, in accordance with the present invention, the values of the pixels defining the image under the cursor are added to LUT-1. The image is assigned to LUT-2, which maps the 8-bit pixels currently defining the image under the cursor to itself. used to. That is, LUT-2 defines 1.1 identical functions. LUT-1 is then assigned to the image currently under the cursor. In this way, the user can determine that the 12-bit pixels of each image that the user has modified correspond to L
You can continue to change the window or level of the remaining image being displayed, which is replaced by the value of UT-1, and that image is assigned to LUT-2.

[Example] Generally, a considerable number of binary bits, for example 12 bits, are
Used to quantize pixels.

This means that a 12-bit pixel specifies one of 4096 intensity levels ranging from black to white in brightness. The 4096 intensity levels consist of what is commonly referred to as a 12-bit gray scale.

However, many display devices use an 8-bit gray scale as described above. Differences between 8-bit and 12-bit grayscales are generally handled by converting the 12-bit format to 8-bit using a transformation mapping as shown in FIG.

In particular, FIG. 1 illustrates four transform maps, A through D, each defined by a respective dynamic range and level. The term "dynamic range", hereinafter referred to as "window", will here mean the width of the occupied portion of the gray scale. For example, among the four transform maps, the window of map "A" occupies the largest portion of the gray scale, ie, the distance between point "a" and point "h".

As mentioned above, transformation mappings are also defined by levels. Here, level refers to the value of a 12-bit pixel mapped to the midpoint of an 8-bit gray scale, ie, a point at gray scale level 127. Therefore, the level of mapping "A" is a pixel with a value of 2048. This is indicated by a "d" and a dashed line. Also, the mapping “
B “C” and “D” levels are each 1750.17
It is 50.2750. (The latter value is designated as "f"). Mapping "C" represents the case where the window of mapping "B" is changed but the level is not changed, which causes a corresponding change in the contrast of the displayed image associated with f1. Mapping "D" represents the case where the level of mapping "B" is changed but the window is not changed, which causes a corresponding change in the intensity of the associated image.

The method of defining the window and level values or respective transformation mappings will be described in detail below. However, for now it suffices to say that the transformation map is expressed by a general formula of the form 0 (x, y) −M [I (x, y)
y)] (1,) where I (x, y) is the value of the input at image coordinates (x, y) (e.g. 1
2-bit pixel value), and O(x, y) is the output value at image coordinates (x, y) (e.g., 8
bit pixel value), and M is a linear mapping image function. However, each 12-bit pixel of the image can be calculated using the formula (1
) to convert to the corresponding 8-bit pixel is a time-consuming method. A faster and well-known method is
The value of each 12-bit pixel is used to select the address of a so-called look-up table (LUT) to obtain the corresponding 8-bit gray scale value of the pixel.

In particular, for each mapping from “A” to “D”, each 1
It can be seen that the value of a 2-bit pixel maps to one of the 256 Gray]0 Gray scale values. Therefore,
each 12 using a predetermined mapping, e.g.
It is a simple matter to calculate the 8-bit gray scale for a bit pixel and store the result in a LUT memory defined by the value of that 12-bit pixel. Thereafter, the conversion function will involve simple table lookup operations.

In teleradiology systems, such as AT&T's CommVjev system, a significant number, e.g.
It often happens that several images are displayed simultaneously on a display device. - Once a degree image is displayed, a user, e.g., a radiologist, may change the window or level of one or more of the displayed images to facilitate diagnosis. For this reason, such a system requires several L
Use UTs, for example 4 LUTs. These will be referred to as LUT-0 to LUT-3, respectively.

Additionally, such a system adds two bits to each pixel of the image to select the address of one of the four LUTs. In other words, 2 of 00,01.1o111 represented by 2 bits
The hexadecimal values are used to select each address of one of the four LUTs.

Initially, the 2-bit value added to each pixel of the image that will eventually be displayed is set to one of the binary values, eg, 00, representing LUT-0. Furthermore, the corresponding 8-bit grayscale value stored in LUT-0 represents a so-called default value derived from a predetermined transformation function as represented by mapping "A". Therefore, the 20 images have the same window and level values,
That is, they are displayed as values of 4096 and 2o48, respectively. Thereafter, the user may change the window or level values for any of the displayed images.

Typically, when a user wants to change the window or level of an image, they move the screen cursor over the image and issue commands to change the window or level of the image under the cursor through an input device such as a terminal or trackball. Enter into the system.

With a trackball, the user rotates the trackball about its respective axis to change the window or level of the image. That is, the teleradiology system changes the window of the image under the cursor in proportion to the increasing rotation of the l-rack ball about the horizontal axis. Similarly, the system changes the level of its image in proportion to increasing rotation of the trackball about its vertical axis.

When the user starts changing the window or level of the image under the cursor, the system responds by changing the 2-bit LUT address of each pixel in the image from address 00 (LUT-0 to address 01 (LUT
-1). The system then transfers a copy of the 8-bit pixel value stored in LUT-0 to LUT-0.
Load into 1.

Next time the user changes the window or level of that image, the system retransforms the 8-bit pixel values stored in LUT-1 according to the transformation function 3 defined by the current values of window and level. or remap. The system continues this process as long as the user changes the window or level of the image under the cursor.

Once the user is satisfied with the intensity and contrast of the image under the cursor, the user may move the cursor to another displayed image and change the window or level of that image. However, in this case, when the user begins changing the window or level of that image, it is assigned to the pixel LUT-2 that defines the image currently under the cursor. Thereafter, in accordance with the method described above, the system remaps each 12-bit pixel value into an 8-bit value corresponding to the direction in which the user moves the trackball and stores the results in LUT-2.

As can be appreciated from the foregoing, such conventional systems typically use only four LUTs, allowing the user to change only two windows or levels of the displayed image. It is restricted so that it cannot be done. Alternatively, such conventional systems could be trained to allocate one LUT for each displayed image. However, this means that such conventional systems require, for example, 20 LUTs, one LUT for each displayed image if 20 images are simultaneously displayed on the display device. means. Additionally, the 2 bit LUT address needs to be increased to 5 bits to provide only one address for each of the 20 LUTs.

The device of the present invention eliminates the need for 20 LUTs. In fact, the device of the present invention uses only three LUTs, yet allows the user to independently change the window and level of each displayed image. In particular, as done in conventional devices, LUT-0
It is used to map 2-bit pixels to 8-bit pixels as a function of a respective transform mapping, eg, transform map "A". From then on, when the user begins to change the window or level of the image under the cursor, the two address bits associated with each 12-bit pixel of that image represent LUT-1. 1 and a copy of the contents of LUT-0 is read into the LUT-0.As mentioned above, LUT-1 is then changed to
Used to map 2-bit pixel values to 8-bit values.

However, where the present invention differs from conventional devices is when the user subsequently changes the window or level of another image.

When the user so changes the window or level of another image, according to the present invention, the device of the present invention (a) changes the window or level of the image under the cursor to 12 Each bit pixel is LU
(b) reassign the image from LUT-1 to LUT-2; At that time, LUT-2 is 1:
1. are set to be the same function. That is, L
The UT-2 is configured to map the 8-bit gray level pixels that currently make up the image to itself when the image is actually being displayed on the display. Additionally, the window and level values that define the mapping function of LUT-1 are stored in memory. LUT-1 is then assigned to the image currently under the cursor. (The reason for preserving such window and level values will become clear below.) For example, the final transformation of the 8-bit pixel contained in LUT-1 is the mapping "B" shown in Figure 1. Assume that it corresponds to a linear function of . Then, in the device of the present invention, (a
) 0 to 1250 (displayed as “C”).
Replace the 2-bit pixels with 8-bit pixels with a value of 0, and (b) replace each 12-bit pixel with a value in the range 1.251 to 2250 (denoted as "e") with the dynamic range of mapping "B". (c) Replace 12-bit pixels with values in the range 2251 to 4096 with 8-bit pixels having a value of 255. Furthermore, pixel values from 7 O to 255 are stored in respective memory locations of LUT-2 so that when the image is actually displayed on the display device, LUT2 acts as the same function as described in 1.1 above. Ru.

In summary, in the apparatus of the present invention, LUT-0 has default values that are transformed according to the linear function given by mapping "A" in FIG. Each image is then assigned to LUT-0 when the image is displayed on the display. LUT-1 is then assigned to the image under the cursor when the user starts changing the window or level of that image, and LUT-2, which contains 1:1 identical functions, is assigned to the image whose window or level has been changed. Assigned.

Therefore, the user can change the window or level of each displayed image, but in that case, all images will be assigned to LUT-2, except for the image currently under the cursor, which will be assigned to LUT-1. It will be done. However, if the user moves the cursor away from the image in LUT-1 and positions it on another image, the method described above will cause the user to change the window or level of the image currently under the cursor. The image that was under the cursor up to 8 is reassigned to LUT-2. moreover,
In accordance with aspects of the present invention, the LUT-1 conversion function that existed when the image currently under the cursor was last reassigned from LUT-1 to LUT-2 is restored. This is done as follows.

(a) replaces the 8-bit grayscale pixels that make up the image currently under the cursor with the first of the corresponding 12-bit values; (b) reassigns the image from LUT-2 to LUT-1; c) Copy the contents of LUT-0 to LUT-1. In this way, first, when the image is redisplayed during the next so-called refresh period, the window and level of the image under the cursor are restored to their initial, default values. However, in accordance with another aspect of the invention, the contents of LUT-1 are then changed according to the window and level values that were previously saved when the image was last reassigned from LUT-1 to LUT-2. will be reconverted.

For example, suppose the values stored as window and level are 500 and 2500, respectively. As mentioned above, the term "level" refers to the midpoint of the dynamic range (window) of the transform mapping; that is, a value of 2500 maps to an 8-bit pixel, or a value of 127 in grayscale. It also represents the midpoint of the window.Thus, for the assumed values, the window starts at 2250 (2500-1/2 500) and ends at 2750 (2500+1/2 500), as shown in FIG. In this way, by "knowing" the values of the windows and levels, the present invention is able to calculate a linear function, denoted by the mapping "E", and retransform the image itself according to LUT=1. Is it possible to do this?

We now turn to a description of the hardware implementing the invention, as shown in FIG.

In particular, FIG. 3 shows a simplified block diagram of a so-called image recording and communication system 10. As shown in FIG.

Central processing unit 100 performs system control, host processing, and memory management, among other things. System control functions include, for example, system clock generation, interrupt control, and bus determination. Central processing unit 100 communicates with peripheral circuitry via bus 105 . These peripherals include memory 110, hard disk 115, data acquisition management system (DMS) 130, and graphics system processing unit (GSP) 135. The central processing unit 110 is
- other peripherals, collectively designated 125, as well as bus 1
05 and I10 interface circuits 120. These other peripherals include, for example, a pig entry terminal, a maintenance port 1, and a printer.

DMS 130 consists of a host processing unit and at least one large capacity disk (not shown). This disc is, for example, an optical disc. DMS 13
0 includes a data acquisition circuit, such as a so-called "frame grabber" circuit, for obtaining a digital version of the image output by the scanner 170 on the bus 171. That is, the frame grabber circuit:
[Scanner] 70 or 1 "Grabs" the analog signal output to bus 171 (gr
abs), converting the analog signal into digital picture elements (pixels) for storage on the DMS 130 disk.

From then on, users such as radiologists can use DM
System 10 will be requested to display the image stored in S 130 on display device 160. To do this, the user enters the identification number associated with the image (=J) through the terminal 185. In response to receiving the identification number over the bus 166, the central processing unit
A command including that number is sent to the DMS 130 through 05.

In response to the instructions, DMS 130 reads the 12-bit pixels of each image in the series of images identified by its number from the optical disk and downloads them, line by line, onto bus 105.

Central processing unit 100 then causes GSP 135 to remove from bus 105 the 12-bit pixels associated with the first group of images in the series. GSP 135 includes, among other things, a display processing unit and memory, such as 2 so-called DRAM memories (not shown). GSP135 is a 2-bit pixel
Stored in RAM memory, 12-bit pixels are stored on bus 10
removed from 5. At this time, graphics system processor 135 adds the aforementioned 2-bit LUT address bits to each 12-bit pixel. At this time, the address bit represents LUT-0. The copy of the image stored in the memory of graphics system processing unit 135 is ultimately transferred to frame buffer 1 via local bus 136.
45.

Subsequently, the image stored in frame buffer 145 is displayed on display device 160.

As mentioned above, each 12-bit pixel is associated with a 2-bit address τj that identifies each of LUT-0 to LUT-2. As you can see from the figure, memory 1
50 is 4 LUTs, i.e. LUT-0 to LU”
It is divided into up to l-3.

That is, the memory 150, for example, is divided into four parts, each part consisting of 40963 (41) memory locations, each memory location being 8 bits. occupies memory locations 0 to 0 to 4095, L
UT-1 occupies memory locations 4096 through 8191, and so on. In carrying out the invention,
LUT-3 is reserved for future use. 12
Bit pixels and their associated 2-bit LUT addresses are thus mapped to specific locations in memory 150, where each 8-bit converted value is stored.

The 8-bit pixel conversion value is then provided over bus 151 to digital-to-analog conversion (DAC) circuitry 155, which converts the pixel to an analog signal.

Converter 155 then provides the results to display device 160.

System 10 includes a control panel 165. This control panel incorporates, among other things, a microprocessor control interface for controlling the operation of panel 165.

4 The transmission of data representing an event, such as an increase in rotation of the trackball 167 in a particular direction, is reported to the DSP 1.35 over the bus 166.

Transmission of signals over bus 166 occurs as an R3232 or TTI serial data stream.

Once the images are displayed on the display 160, as shown in FIG. Changes may be initiated. Once the user is satisfied with the window or level of that image, the user may move the screen cursor to another image, such as image 17, and subsequently change that image. In fact, as mentioned above, a user can change the windows and levels of the remaining displayed images using only three LUTs in accordance with the present invention.

Here, the present invention is applied to the graphics system processing unit 135.
Let's move on to the explanation of the software that will be executed.

5 In particular, when the screen cursor 200 is positioned at a certain image, for example image Ill, and the user begins to rotate the trackball 167 in a particular direction, the fifth
The illustrated program begins at block 500. From block 500, the program starts at block 5
Go to 01 and the image under the cursor at the moment is LUT-
1. If, in fact, the image currently under the cursor is assigned to LUT-1, the program moves to block 505. Otherwise, the program proceeds to block 502 where it checks to see if another display image is assigned to LUT-1. If another image is actually assigned to LUT-1, the program moves to block 507. In other cases,
The program proceeds to block 503 and checks whether the image currently under cursor 200 is assigned to LUT-2. If in fact it is currently under cursor 200, the program moves to block 508. Otherwise, proceed to 504.

6. At block 504, the program assigns the image to LUT-1 after copying the contents of LU"r-0 to LUT-1 in the manner previously described.

The program then proceeds to block 505. Block 5
At 05, the program retransforms LUT-1 according to the last increment of window and level received from trackball 167 over bus 166. The program then passes through block 506 and ends.

As mentioned above, if the program determines that another image is assigned to LUT-1, it moves to block 507. At block 507, the program reassigns other images from LUT-1 to LUT-2 in the manner previously described and continues to block 503.

As mentioned above, the program currently has cursor 200
If the underlying image is assigned to LUT-2, the process moves to block 508. At block 508, the program (a) recreates the 8-bit transformation function that was stored in LUT-1 when the image was last reallocated from LUT-1 to LUT-2; and (b) After reading the data of the first image from the memory of the GPS 135, it is restored by loading it into the frame buffer] 45, and (C) the image is transferred from LUT-2 to
Reallocate to UT-1.

The program then proceeds to block 505 when the foregoing tasks are completed.

Having shown and described specific embodiments of the invention,
It will be understood that various modifications may be made without departing from the scope and spirit of the invention.

For example, the present invention can be used to display a plurality of images on a display device, and any type of image window or level in which the user can change the window or level of any image in the displayed images. Can be readily applied to any type of imaging system. Furthermore, in the present invention, a case has been described in which a 12-bit pixel is converted into a corresponding 8-bit pixel, but the present invention also provides a method for converting a 12-bit pixel into a corresponding 8-bit pixel.
It is equally effective in the case where it is converted or mapped into a bit pixel (although M may be larger than N and may be smaller than 8).

[Brief explanation of drawings]

FIG. 1 shows some possible transformation mappings useful for understanding the concept of windows and levels with respect to image processing; FIG. 2 shows another possible transformation mapping; 3 is a schematic block diagram of a so-called teleradiology system in which the present invention is conveniently implemented; FIG. 4 is a diagram showing the display device of FIG. 3 with a plurality of images displayed; FIG. 4 shows a flow diagram of a program for controlling the operation of the graphics system processing unit of FIG. 3 in accordance with the principles of the present invention. Applicant: American Telephone & Fl (3
,4 Io,5

Claims (5)

[Claims] (1) means for displaying a plurality of images of pixels on a display according to a general transformation function defined by predetermined window and level values; and the values of at least one window of the displayed images are changed. means for causing said one image to be redisplayed according to another transformation function defined by the current window and level values of said one image in response to said one other window of said displayed image; means for replacing a set of pixels defining said one image with a value corresponding to said set of pixels derived from said other transformation function in response to at least a change in a value of said one image; 1
An image display device comprising means for re-displaying according to a conversion function defined by a pair-to-one identical function. (2) The general conversion function and the other conversion function are linear conversion functions, the slopes of which are defined by respective window and level values. Image display device. (3) The replacing means includes means for storing in a memory the window and level values of the one image at the time when the pixel group of the one image is replaced. Image display device. (4) restoring the pixel group of the one image if the window value of the one image at least starts to change after the window value of the other image is changed; 2. An image display device according to claim 1, further comprising means for redisplaying said one image as a function of a transformation function derived from window and level values stored in said image. (5) An image display method for independently controlling at least one of the windows and levels of each of a plurality of displayed images, comprising the steps of: storing a group of pixels defining each of the plurality of images; , transforming according to a general transformation function defined by predetermined window and level values and displaying the result on a display; accordingly, modifying a copy of said general transformation function according to said modified values, retransforming said one image according to said modified transformation function, and displaying the result on said display as said one image. deriving each set of pixels defining said one image from said modified transformation function in response to at least one of the window, level for another one of said displayed images being changed; and subsequently displaying said derived pixels as said one image using a one-to-one conversion function.