JP3019735B2 - Image processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for increasing the number of image matrices by interpolation processing, and more particularly to an image processing apparatus suitable for processing a digital fluoroscopic image obtained by a fluoroscopic apparatus. Related to the device.

[0002]

2. Description of the Related Art In an X-ray fluoroscopic apparatus, a digital X-ray fluoroscopic image is usually obtained as an interlaced image. In order to display such a digital X-ray fluoroscopic image on a high definition monitor device or to enlarge an image, the number of image matrices may be increased by interpolation processing. In such a case, conventionally, an interlaced image is generally treated as a non-interlaced image, and new image data located between scanning lines of the original image is generally created by interpolation processing.

For example, when the number of matrices is doubled in the scanning line arrangement direction and linear interpolation is performed as an interpolation method, the following is performed. The original image is represented by each scanning line as shown in FIG. As shown in FIG. 5A, in the odd field and the even field, the position of the scanning line is located between the other scanning lines. As shown in FIG. 5B, the interpolated image of this image has, in the odd field, each scanning line of the original even field between the scanning lines O1, O2, O3,. It is made by positioning E1, E2, E3,. In the even field, a scanning line located between each scanning line of the newly created odd field is obtained by interpolation processing. That is, for example, the original odd-field first scanning line O1 and the original even-field first scanning line E1 are added to add 2
Is obtained, and is displayed at an intermediate position between the first scanning line O1 of the original odd field and the first scanning line E1 of the original even field.

Alternatively, interpolation of only data in a field as shown in FIG. 5C is also performed. here,
The odd field after interpolation is created using only the scan lines of the original odd field, and the even field after interpolation is created using only the scan lines of the original even field. In the interpolated odd field, the scanning line of the original odd field is used as it is and displayed at the original position, and a new scanning line is interpolated between the original scanning lines. In the even field after the interpolation, if the same interpolation as that of the odd field is performed, the position of the scanning line overlaps with the scanning line of the odd field. For example, the distance between the scanning lines E1 and E2 of the original even field is set. Scan lines to be positioned (1/4) above (2E1 + E
2) / 2, and a scanning line to be positioned (（) below the interval is obtained by (E1 + 2E2) / 2.

[0005]

However, according to the conventional interpolation method, there are problems such as horizontal streaks due to the movement of the image, and the high-frequency component of the image is reduced and the image is blurred. . That is, in the interpolation as shown in FIG. 5B, the original odd-field scanning lines and the even-field scanning lines appear alternately in the odd field. Because the original oddfield and evenfield have a time difference, of course,
Especially when displaying a moving image, if there is a large movement between the odd field and the even field, considerably different data will be displayed for each scanning line, which causes streaks in the horizontal direction. The image looks as if it had entered. On the other hand, in the case of FIG.
Although the data displayed in each field after interpolation does not include data having a time difference, high-frequency components of an image are attenuated because interpolation is performed using scanning lines that are far apart.

SUMMARY OF THE INVENTION In view of the above, it is an object of the present invention to provide an image processing apparatus improved so that an appropriate image interpolation process can be performed according to the motion of an image.

[0007]

In order to achieve the above object, in an image processing apparatus according to the present invention, a first interpolation for performing different interpolation in accordance with the motion of a digital X-ray fluoroscopic image input as an interlaced image. Interpolating means and second
Is selected.

[0008] The first interpolation means performs interpolation using only data in one of the odd and even fields of the digital X-ray fluoroscopic image input as an interlaced image, and performs the other interpolation in the other field. The interpolation is performed by using the data in the other field and the data in the one field, and the second interpolating means performs the interpolation on the data in one of the two fields and the data in the other field. Interpolation is performed using the data of the other field.

The switching means for switching between the first and second interpolation means switches to the first interpolation means when the input image has a motion, and the input image has no motion. In some cases, the operation is switched to the second interpolation means.

[0010]

The first and second interpolation means for performing different interpolation are switched depending on whether the digital X-ray fluoroscopic image input as an interlaced image is a moving image or not. That is, when the input image is a moving image, the operation is switched to the first interpolation unit. When the input image is a motionless image, the operation is switched to the second interpolation unit.

[0011] The first interpolation means used when the digital X-ray fluoroscopic image input as an interlaced image is a moving image, uses only the data in the odd or even field in one of the fields. Interpolation is performed, and in the other field, interpolation is performed using the data of the other field and the data of the above-mentioned one field, so data having a time difference enters only the latter field. Disturbance can be suppressed. In the latter field, data of two fields having a time difference is used. However, since interpolation can be performed using data at a closer position, attenuation of high frequency components of an image can be reduced.

[0012] The second interpolation means used when the digital X-ray fluoroscopic image input as an interlaced image is a motionless image, includes one of both the odd and even fields, and one of the two fields. Since the interpolation is performed using the data of the field and the data of the other field, the data of the other field is fetched in any field. In this case, since there is no motion in the image, the time difference between the fields is a problem. Instead, a better interpolated image can be obtained.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a preferred embodiment of the present invention will be described in detail with reference to the drawings. In FIG. 1, an interlace scanning image signal is sent to an input terminal IN. That is, as shown in FIG. 2A, odd fields FO1, FO2, FO3, FO4,... And even fields FE1, FE2, FE3, FE4,.

The field memory 11 stores only the odd-field image signal, and is written in the odd-field period (for example, FO1), and in the next even-field period (FE1) and the following odd-field period (FO2). This stored FO
One image signal is read. Therefore, as shown in FIG. 2B, the output A of the field memory 11 is delayed by one field with respect to the input, and the image signal of the same odd field is repeatedly output for a period of two fields (one frame period). Will be. In a field period after outputting the same odd-field image signal, writing of the next odd-field image signal and reading of the previous (stored) odd-field image signal are performed simultaneously. Such control of writing and reading is performed by the controller 23.

The field memory 12 stores only an even-field image signal, and is written in an even-field period (for example, FE1), and is read out in the next odd-field period (FO2). . The write / read control is performed by the controller 23. A switching circuit 13 is provided on the output side of the field memory 12, and the input side a and the output side b of the field memory 12 are shifted to the a side in the even field period as shown in FIG. Switch to the b side during the odd field period. The control of the switching circuit 13 is also performed by the controller 23.

As shown in FIG. 2C, at the point B, the image signal of the even field appears as it is in the even field period of the input, and the image signal of this even field appears in the next field period (odd). In the field period). In other words, the field memory 12 cannot read the written image signal at the same time as writing the image signal of the even field during the even field period. Switching circuit 13 without intervention of field memory 12
Is output from the field memory 12 to the point B via the side a of the switching circuit 13 during the next odd field. It is trying to make it.

The switching circuit 14 is controlled by the controller 23 so as to be switched to the side a in the even field of input and to the side b in the odd field as shown in FIG. 2E. Therefore, at point C on the output side of the switching circuit 14, as shown in FIG. 2 (f), the image signal of the immediately preceding odd field read from the field memory 11 during the input even field period appears. In the next field period (odd field period), the image signal of the immediately preceding even field read from the field memory 12 appears. Therefore, at this point, the input image signal appears as it is with a delay of one field.

The image signal at point C is input to the line memory 15, and a signal delayed by one scanning line and a signal delayed by two scanning lines are obtained. The line memory 16 obtains a signal delayed by one scanning line, and receives the output A of the field memory 11.

The arithmetic units 17 and 18 are for lower line interpolation and upper line interpolation, respectively, and newly calculate the image signals on the four scanning lines simultaneously inputted to the four input terminals 71 to 74 and 81 to 84, respectively. A good image signal. That is, a new image signal is created by weighting and adding the input signals of the four lines, and the weight coefficient is set by the controller 23. The output P of the arithmetic unit 17 for lower line interpolation is switched by the switching circuit 19 to a signal output from the line memory 15 and delayed by one scanning line. The same applies to the output Q of the arithmetic unit 18 for upper line interpolation. The switching circuit 20 switches the signal output from the line memory 16 to a signal delayed by one scanning line.

The switching circuits 19 and 20 are controlled by a controller 23. The switching circuit 19 is always on the b side as shown in FIG. for that reason,
The output P of the lower line interpolation arithmetic unit 17 always appears at the output d of the switching circuit 19. As a result, as shown in FIG. 2 (i), this output d has a lower line interpolation arithmetic unit 17 for obtaining an interpolated lower line output of an input odd field and an output odd field.
.. Appear, and output PE1, PE2, PE3,... Of the lower-line interpolation arithmetic unit 17 for obtaining an interpolated lower-field output of an output odd-field in an input odd-field. Appears.

On the other hand, as shown in FIG. 2H, the switching circuit 20 is switched to the a side in the even field of the input and to the b side in the odd field. Therefore, as shown in (j) of FIG.
In the input even field, the field memory 11
FO1, FO2, FO3, FO3, FO2, FO3, which are obtained by delaying the image signal of the immediately preceding odd field read from
... appears. This image signal is a signal that becomes an upper line output of the odd field of the output as described later. In the input odd field, the outputs QE1, QE2, QE3,... Of the upper line interpolation arithmetic unit 17 appear as an interpolated upper line output of the output even field.

The image signals appearing at the outputs D and E of the switching circuits 19 and 20 are simultaneously written to the field memories 21 and 22 line by line, and are alternately read line by line with a delay of several lines. These write / read controls are performed by the controller 23. Writing to the field memories 21 and 22 is performed simultaneously and in parallel. On the other hand, the read time for one line is set to half the time of the scan line in the original field, and the field memories 21 and 22 are read.
Are read alternately line by line. As a result, in each of the original field periods, an image signal of twice as many scanning lines is output.

Further description will be made with reference to FIG. As shown in FIG. 3A, an input odd-field image is composed of scanning lines O1, O2, O3, O4, O5,.
3, E4, E5,... In this case, the input even field FE shown in FIG.
Considering the period 1, the image signal of the immediately preceding odd field appears at the point E with a delay of one scanning line, and this is written into the field memory 22 line by line. Therefore, this field memory 22 has the scanning lines O1, O2, O3, O4 on the left side of FIG.
The signals on O5,... Are held.

On the other hand, in the period of the input even field FE1 shown in FIG. 2A, the image signal PO1 from the lower line interpolation arithmetic unit 17 appears at two points. In the image signal PO1, the signals on the respective scanning lines are generated from the signals on the scanning lines input to the input terminals 71 to 74. At some point, input terminal 72
Assuming that the current scanning line O3 of the odd field FO1 appearing at the point A has been input, the input terminal 7
4, a signal obtained by delaying the image signal of the odd field FO1 appearing at the point A by one line in the line memory 16, that is, the scanning line O2 one line before in the odd field FO1 is input. During the period of the even field FE1 of the input, the same signal as the point A, that is, the image signal of the odd field FO1, appears at the point C, so that the input terminal 71 of the arithmetic unit 17 has the same odd field FO1 as the input terminal 72. Current scan line O
3 is input. The input terminal 73 receives the image signal of the point C delayed by one line via the line memory 15, so that the input terminal 73 has the same scanning line O 2 as the previous line in the odd field FO 1.
Is entered.

In this arithmetic unit 17, the input terminals 72, 7
A weighting factor of 0.5625 for the input to 3 and (-0.0625) for the input to the input terminals 71 and 74
, And are added after being given. Therefore, the second scanning line output from the arithmetic unit 17 at this time is formed by 0.5 × O2 + 0.5 × O3. Since the other scanning lines are the same,
(O1 + O2) / 2, (O2 + O3) / 2, (O3 + O
4) / 2, (O4 + O5) / 2,... This is written and held in the field memory 21.

First, all the signals O1 of the first scanning line are read from the field memory 22 in half the read time, and then all of the signals (O1 + O2) / 2 of the first scanning line are read from the field memory 21. Since the readout is performed in half the readout time and this is sequentially repeated, the image signal of the odd field for this output is used as it is, as shown on the left side of FIG. , A scanning line formed from two original odd-field scanning lines located immediately before and after the intermediate scanning line of the original odd-field.

Next, during the period of the input odd field FO2 shown in FIG. 2A, the image signals PE1 and QE1 from the arithmetic units 17 and 18 appear at points D and E, respectively. The signal at point C and a signal obtained by delaying the signal by one line are input to input terminals 71 and 73 of the arithmetic unit 17, and the signal at point C is an even field signal as shown in FIG. At this point, if the signals of the input terminals 72 and 74 are the signal of the third line O3 and the signal of the second line O2 of the odd field FO1, the input terminal 7
Reference numerals 1 and 73 are the signal of the third line E3 and the signal of the second line E2 of the even field FE1. Therefore, the signal of the second line output from the arithmetic unit 17 at this point is (−0.0625) × O2 + 0.5625 × E2 + 0.
5625 × O3 + (− 0.0625) × E3. Other scan lines are made in the same way,
If the scanning line with a small coefficient at the end is omitted and the coefficient 0.5625 is rounded to 0.5 and expressed, in order:
(E1 + O2) / 2, (E2 + O3) / 2, (E3 + O
4) / 2, (E4 + O5) / 2,..., Which are written and held in the field memory 21.

The input terminal 81 of the arithmetic unit 18 receives the signal of the current line O3 of the image signal appearing at the point A, and the input terminal 83 delays the signal by one line in the line memory 16, that is, the line O2. Is input. The input terminal 82 receives 1 from the line memory 15.
Since the line delay output is input, the signal of the line E2 of the even field FE1 is input, and the input terminal 8
4, the signal of the even field FE1 delayed by two lines by the line memory 15, that is, the signal of the line E1 is input.

In the arithmetic unit 18 as well, the weighting coefficient is 0.56 for the input to the input terminals 82 and 83.
25, the input to the input terminals 81 and 84 is (−0.
0625), the arithmetic unit 18 at this time.
Output from the second line is (−0.06
25) × E1 + 0.5625 × O2 + 0.5625 × E
Made by 2 + (− 0.0625) × O3. The other scanning lines are formed in the same manner. The scanning lines having small coefficients at the ends are omitted and the coefficient
(O1 + E1) / 2, (O
2 + E2) / 2, (O3 + E3) / 2, (O4 + E4)
.., And this is written and held in the field memory 22.

In reading from these field memories 21 and 22, first, all signals of the first scanning line (O1 + E1) / 2 are read from the field memory 22 in half the read time, and then read from the field memory 21. All of the signals (E1 + O2) / 2 of the first scanning line are read out in half the readout time, and this is sequentially repeated. Then, an output image of the even field is formed as shown on the right side of FIG. 4B by an image signal whose scanning lines are doubled by such reading. If the position of each scanning line is arranged between the scanning lines of the odd field, a signal appropriate as a signal of the scanning line at that position is interpolated. That is, for example, the third line of the output image of this even field is (O2 + E2) / 2, which is interpolated between the line O2 of the input odd field and the line E2 of the input even field. It is appropriate as a signal to be used, and the fourth
The second line is (E2 + O3) / 2, which is suitable as a signal to be interpolated between the input even-field line E2 and the input odd-field line O3.

As described above, as shown in FIG.
In the odd field, double scanning lines are obtained by interpolating using only the original odd field data, and in the even field, double scanning is performed by interpolating using both the original odd and even field data. It is a line. Therefore, since only data having no temporal difference is used in the odd field, it is possible to prevent the occurrence of striped artifacts in the scanning line direction. On the other hand, in the even field, interpolation is performed using data of both the odd and even fields, so that data of a scanning line at a close position is used instead of data of a scanning line at a distant position. ,
It is possible to prevent a high-frequency component from being lost in the scanning line arrangement direction and the image from being blurred. Of course, in the odd field, high-frequency components are lost because interpolation is performed using data of distant scanning lines, and in the even field, there is a problem of using data with a time difference. By combining these images, those problems are alleviated, and an overall excellent image is obtained.

Further, when there is no motion in the image, the switching circuit 14 is always switched to the b side as shown in FIG.
As shown in (g), the controller 23 may control to alternately switch between the a-side and the b-side for each field as in the switching circuit 20. Others are the same as those in FIG. In this case, the input odd fields FO2, FO3, FO4,..., That is, the output even field are the same as those in FIG. 2, and therefore each scanning line of the output even field image is shown in FIG. Like the right side, it is the same as the right side in FIG.

On the other hand, in the input even fields FE1, FE2, FE3, FE4,..., Since the switching circuit 14 is always on the b side, the point C in FIG.
(F), the even fields FE1, F
The image signals of E2, FE3, FE4,... Appear. Therefore, the signal of the current line of the even field and the signal of the previous line are input to the input terminals 71 and 73 of the arithmetic unit 17, respectively.
Signals one line before and two lines before the even field are input to 2 and 84, respectively. At this time, the signal of the current line of the immediately preceding field, that is, the image of the odd field constituting the frame is input to the input terminal 72 of the arithmetic unit 17 and the input terminal 82 of the arithmetic unit 18. Apparatus 1
7 and the input terminal 84 of the arithmetic unit 18 are supplied with a signal one line before the image of the odd field.

However, in the input even fields FE1, FE2, FE3, FE4,..., The switching circuits 19 and 20 are all on the a side as shown in FIGS. 3 (g) and (h). The outputs of the arithmetic units 17 and 18 are not used, and the two fields have even fields FE1, FE2, FE3, FE4 and FE4 as shown in FIG.
.. Appear with a delay of one line.
(J), the immediately preceding odd fields FO1, FO2, and FO output from the field memory 11 are output.
, FO4,... Appear one line later. The even-field and odd-field image signals are written into the field memories 21 and 22, respectively, and are read alternately one line at a time at double speed, with a delay of several lines after the writing. As shown in (c), the original odd-field lines O1, O2, O3, O4, O5,... And the original even-field lines E1, E2, E3, E
4, E5,... Appear alternately.

Therefore, in this case, in the odd field, since the interpolation is performed by using each line of the original even field as it is, a decrease in the high frequency component due to the interpolation from the data of the line which is distant from the position is avoided. be able to. Since such interpolation is performed when there is no motion of an image such as a still image, there is no problem caused by the fact that data having a temporal difference exists in the odd field.

The mode switching between the interpolation mode shown in FIG. 4B and the interpolation mode shown in FIG. 4C is performed according to whether the input image is a moving image or a still image. This mode switching can be automatically performed when applied to an X-ray fluoroscopic apparatus. In other words, a control signal from the X-ray fluoroscope is input to the controller 23, and during fluoroscopy, the interpolation mode shown in FIG. 4B is entered, and during film photography, the interpolation mode shown in FIG. Switch automatically. At the time of fluoroscopy, since the X-ray image signal is being sent in real time one after another, it is a moving image,
Therefore, by setting the interpolation mode shown in FIG. 4B, it is possible to suppress image disturbance due to a time difference between fields. On the other hand, during film shooting, the image signal of the last frame of the fluoroscopy immediately before that is repeatedly output as a last image hold image. For this reason, since a still image signal is output, the time difference between the fields is not a problem. Therefore, by setting the interpolation mode of FIG. 4C, a more excellent interpolated image can be obtained. .

In the above embodiment, one line is interpolated using four lines. However, as shown in the abbreviation above, an interpolated line can be formed from two lines. . Also, the controller 23 controls each switching circuit and changes the weighting coefficients of the arithmetic units 17 and 18 so as to set the interpolation mode of FIG. 4C for a still image and to set the interpolation mode of FIG. ), The interpolation can be performed using only the data in the original fields.

[0038]

As described above, according to the image processing apparatus of the present invention, when there is a motion in an image, interpolation is performed so as to suppress disturbance of the image due to a time difference between fields. When there is no motion, it is assumed that there is no time difference between the fields, and an ideal interpolation is performed in terms of position to obtain an excellent image without attenuation of high-frequency components of the image (image blurring). .

[Brief description of the drawings]

FIG. 1 is a block diagram of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a time chart showing signals of respective units in one interpolation mode of the embodiment.

FIG. 3 is a time chart showing signals of respective units in another interpolation mode of the embodiment.

FIG. 4 is a schematic diagram showing scanning lines of an image in the embodiment.

FIG. 5 is a schematic diagram showing scanning lines of an image in a conventional example.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 11 Field memory for input odd field 12 Field memory for input even field 13, 14, 19, 20 Switching circuit 15, 16 Line memory for delay 17 Bottom line interpolation operation device 18 Top line interpolation operation device 21 Lower line after interpolation Field memory 22 Upper line field memory after interpolation 23 Controller

Claims (1)

(57) [Claims] In one of odd and even fields of a digital X-ray fluoroscopic image input as an interlaced image, interpolation is performed using only data in the field, and in another field, interpolation is performed in the other field. First interpolating means for performing interpolation using the data and the data of the one field, and using either one of the fields and the data of the other field for any one of the two fields. Switching to the second interpolation means for performing interpolation and the first interpolation means when the input image is moving, and the second interpolation means when the input image is not moving. And a switching means for switching to the interpolation means.