1. Field of the Invention
The present invention relates to a technology for emphasizing a contour of an image in an image display device, an image pickup device and the like.
2. Description of the Related Art
A conventional technology of a contour emphasizing process is described referring to
The contour component signal generating circuit 33a executes a process shown in
The image data values D13, D12 and D11 respectively allocated to the pixels P32, P22 and P12 are multiplied by the coefficients k6, k5 and k4, and the contour component signal S1 is generated from the total of the multiplication values. The image data value of the pixel of attention P22 is added to the contour component signal S1 by the adder 33b, and a contour emphasizing signal S2 is thereby outputted.
The status shown
(−1)×80h+2×80h+(−1)×0=80h
The contour emphasizing signal S2, which is obtained from the addition of the contour component signal S1 and the pixel of attention D12, results in the following.
80h+80h=FFh
The addition result is FFh as shown above in order to execute an eight-bit precision clip. The contour emphasizing signal S2 is thus increased at the timing t11.
The contour emphasizing signal S2 results in a large value also at a timing t14 because region data outside the image signal (black level signal) shows a value lower than the image signal, as a result of which the contour emphasizing signal S2 results in a value larger than previously intended (t15 and t16). More specifically, the data other than the image signal (D11 at the timing t11 and D13 at the timing t14) are judged to be black signals. Therefore, these signals (t15 and t16), which result in an unnaturally whitened image as the image signals when the contour emphasis is executed, cannot be used as effective signals. As a result, an actually usable output effective pixel is SA3.
The status shown in
The result of adding the value of the contour component signal S1 and pixel of attention is FFh in order to execute the eight-bit precision clip. In the same manner, the contour emphasizing signal S2 results in a large value at a timing t22 because the black level signal shows a value lower than that of the image signal, as a result of which the contour emphasizing signal S2 results in a value larger than previously intended (t23 and t24). These pixels cannot be used as the effective signals because they result in the unnaturally whitened image as the image signals. The output effective pixel in the case of not using these pixels can be represented by SA4.
Showing the contents of the contour emphasizing process in an image signal chart, signals shown in
In the contour emphasizing process is also available a method in which a difference between the central part and the peripheral part in terms of a resolution of an image pickup lens is counterbalanced so as to obtain a substantially even resolution across an entire footage (for example, see the patent literature 1). However, the method, which utilizes a characteristic of the lens to correct the contour, fails to prevent the number of the pixels decreasing in the contour emphasizing process.
In the conventional technology, when the contour emphasizing process is executed to image data having an input image size by n pixels in an x direction and m pixels in a y direction, an output image size is by (n−2) pixels in the x direction and (m−2) pixels in the y direction. There is no particular problem in the foregoing case if the input image size is by equal to or more than (n+2) pixels in the x direction and equal to or more than (m+2) pixels in the y direction. However, an output format of an image sensor element is increasingly diversified along with an increasing number of pixels of the image sensor element, while a display size of the image signal ranges in a wider variety. To response to the ongoing change, a filtering process such as zooming may be necessarily executed in a pre-stage process. Accordingly, the input image size fails to satisfy the foregoing conditions in some cases, in which case the targeted size of the output image cannot be obtained making it not possible to use the contour emphasizing circuit.
A contour emphasizing circuit according to the present invention comprises:
The input/output switching circuit is preferably adapted to extend an effective signal of a pixel boundary to an ineffective signal region of an adjacent pixel boundary based on a control of the image boundary detecting circuit.
The input/output switching circuit is preferably adapted to switch a read-out line when the image boundary detecting circuit detects an image boundary in a vertical direction and switch a read-out pixel when the image boundary detecting circuit detects an image boundary in a horizontal direction.
The image boundary detecting circuit detects the image boundary portion separating effective and ineffective of the image data to thereby control the input/output switching circuit. The input/output switching circuit is interposed between the line memory circuit and the image data retaining circuit and between the image data retaining circuit and the contour emphasizing signal generating circuit.
The input/output switching circuit, when detecting the image boundary in the horizontal direction, interpolates the image data of a pixel adjacent to a boundary outside an effective pixel region by the image data of a pixel adjacent to the boundary inside the effective pixel region, which is carried out on the boundary of the effective pixel region on both of right and left sides thereof. On the left-side boundary, an ineffective pixel next to a pixel of attention (effective pixel) on the left side is interpolated by the pixel of attention, while an ineffective pixel next to the pixel of attention (effective pixel) on the right side is interpolated by the pixel of attention on the right-side boundary.
The input/output switching circuit, when detecting the image boundary in the vertical direction, interpolates the image data of a line adjacent to the boundary outside the effective pixel region by the image data on a line adjacent to the boundary inside the effective pixel region, which is carried out on the boundary of the effective pixel region on both of upper and lower sides thereof. On the upper-side boundary, an ineffective line next to a line of attention (effective line) on the upper side is interpolated by the line of attention, while an ineffective line next to the line of attention (effective line) on the lower side is interpolated by the line of attention on the lower-side boundary.
More specifically, the ineffective signals (black level signal) immediately adjacent to the effective pixel region in the four directions thereof are interpolated by the effective signals adjacent thereto inside the effective pixel region. When the image signal, in which the ineffective signals immediately adjacent in the four directions are interpolated by the adjacent effective signals inside the effective pixel region, is inputted to the contour emphasizing signal generating circuit, the generation of an abnormal contour emphasizing component is prevented in the effective pixel region. Therefore, the output image contour-emphasized in the same size as the input image size can be generated.
The present invention can be realized without providing the input/output switching circuit.
More specifically, a contour emphasizing circuit according to the present invention comprises:
The filter coefficient is adjusted in the foregoing manner so that the ineffective signals (black level signal) immediately adjacent to the effective pixel region in the four directions thereof are interpolated by the effective signals adjacent thereto in the effective pixel region in the same manner as described earlier. The image signal, in which the ineffective signals immediately adjacent in the four directions are thus interpolated by the adjacent effective signals in the effective pixel region, is subjected to an operation process in the contour emphasizing signal generating circuit. As a result, the generation of the abnormal contour emphasizing component is prevented in the effective pixel region. Therefore, the output image contour-emphasized in the same size as the input image size can be generated.
Further, in another effective contour emphasizing circuit, the image boundary detecting circuit is controlled by an image size adjusting signal inputted from outside. More specifically, the image boundary detecting circuit is adapted to halt its action when the image size adjusting signal indicates an increase of number of pixels by two or an increase of number of lines by two relative to a standard.
The contour emphasizing circuit constituted in the foregoing manner can flexibly respond to such a case in which the image size is different depending on an enlarged magnification such as zooming or an output format of an image sensor element.
The present invention is illustrated be way of example and not limitation in the figures of the accompanying drawings, in which like references indicate similar elements and in which:
Hereinafter, a contour emphasizing circuit according to preferred embodiments of the present invention is described in detail-referring to the drawings.
In the contour emphasizing circuit 7 shown in
An image data retaining circuit (filter) 12 comprises a plurality of delay elements 12a-12f. The respective delay elements 12a-12f delay the image data per pixel which is a minimum unit of the image data.
A contour emphasizing signal generating circuit 13 comprises a contour component signal generating circuit 13a and an adder 13b. The contour emphasizing signal generating circuit 13 executes a coefficient operation with respect to a central pixel and peripheral pixels to thereby generate a contour emphasizing signal S2.
An image boundary detecting circuit 14 detects an image boundary portion separating effective and ineffective of the image data. The image boundary detecting circuit 14 outputs a selector control signal S3 when detecting a signal of a left-corner pixel constituting a left-corner part of the image, outputs a selector control signal S4 when detecting a signal of a right-corner signal constituting a right-corner part of the image, outputs a selector control signal S5 when detecting a signal of an upper-corner pixel constituting an upper-corner part of the image, and outputs a selector control signal S6 when detecting a signal of a lower-corner pixel constituting a lower-corner part of the image.
An input/output switching circuit 15 comprises a plurality of selectors 15a-15h and switches between an output of the line memory circuit 11 with respect to the image data retaining circuit 12 and an output of the image data retaining circuit 12 with respect to the contour emphasizing signal generating circuit 13 in response to the selector control signals S3-S6 outputted by the image boundary detecting circuit 14 as a detection result.
More specifically, the following action is executed in the input/output switching circuit 15 in response to the selector control signals S3-S6. The selector 15h selects an output of the line memory 11b when the selector control signal S6 is inactive, while selecting an output of the line memory 11a when the selector control signal S6 is active. The selector 15g selects the input image signal (image data) when the selector control signal S5 is inactive, while selecting the output of the line memory 11a when the selector control signal S5 is active. The selector 15f selects an input to the delay element 12e when the selector control signal S4 is inactive, while selecting an output of the delay element 12e when the selector control signal S4 is active. The selector 15e selects an output of the delay element 12f when the selector control signal S3 is inactive, while selecting the output of the delay element 12e when the selector control signal S3 is active. The selector 15d selects an input to delay element 12c when the selector control signal S4 is inactive, while selecting an output of the delay element 12c when the selector control signal S4 is active. The selector 15c selects an output of the delay element 12d when the selector control signal S3 is inactive, while selecting the output of the delay element 12c when the selector control signal S3 is active. The selector 15b selects an input to the delay element 12a when the selector control signal S4 is inactive, while selecting an output of the delay element 12a when the selector control signal S4 is active. The selector 15a selects an output of the delay element 12b when the selector control signal S3 is inactive, while selecting the output of the delay element 12a when the selector control signal S3 is active.
The input image signal (image data) is inputted to the line memory circuit 11, and inputted to the contour component signal generating circuit 13a from an input side of the input/output switching circuit 15 via output sides of the image data retaining circuit 12 and the input/output switching circuit 15. In the contour component signal generating circuit 13a, a predetermined operation is executed so as to generate a contour component signal S1. In the adder 13b, an image signal of a pixel of attention and the contour component signal S1 are added so as to generate and output a contour emphasizing signal S2.
The status shown
The image boundary detecting circuit 14 sets the selector control signal S3 to an enable state at the timing t11 when detecting the signal of the left-corner pixel. Accordingly, the selector 15c is switched from the state of selecting the output of the delay element 12d up to then to the state of selecting the output of the delay element 12c. As a result, the image data D11 is interpolated by the image data D12 (see a downward arrow). At that time, the image data D11 is interpolated in such a manner that an ineffective signal is interpolated by the image data D12 of the boundary pixel inside the effective pixel region because the image data D12 corresponds to the boundary of effective pixels.
The image boundary detecting circuit 14 sets the selector control signal S4 to the enable state at a timing t14 when detecting the signal of the right-corner pixel. Accordingly, the selector 15d is switched from the state of selecting the input of the delay element 12c up to then to the state of selecting the output of the delay element 12c. As a result, the image data D13 is interpolated by the image data D12 (see an upward arrow). At that time, the image data D13 is interpolated in such a manner that the ineffective signal is interpolated by the image data D12 of the boundary pixel inside the effective pixel region because the image data D12 corresponds to the boundary of effective pixels.
The image signal (image data) in the foregoing state is as shown in
Then, generating the contour emphasizing signal S2 based on the arrayed coefficients as k4=−1, k5=2 and k6=−1, a high-frequency component is not detected because a signal of the pixel P12 is equivalent to the pixel P22 at the timing t11. Further, the high-frequency component is not detected because a signal of the pixel P32 is equivalent to the pixel P22 at a timing t14. In other words, there is no generation of an abnormal contour emphasizing component in the contour emphasizing signal S2. In the contour-emphasized image signal image data), a region represented by SA1 can be used as the effective signal.
The image signal in the foregoing state is as shown in
The status shown
The image boundary detecting circuit 14 sets the selector control signal S5 to the enable state at the timing t21 when detecting the signal of the upper-corner line. Accordingly, the selector 15g is switched from the state of selecting the input image signal (image data) up to then to the state of selecting the output of the line memory 11a. As a result, the image data D21 is interpolated by the image data D22 (see a downward arrow). At that time, the image data D21 is interpolated in such a manner that the ineffective signal is interpolated by the image data D22 of the boundary pixel inside the effective pixel region because the image data D22 corresponds to the boundary of effective lines.
Further, the image boundary detecting circuit 14 sets the selector control signal S6 to the enable state at a timing t24 when detecting the signal of the lower-corner line. Accordingly, the selector 15h is switched from the state of selecting the output of the line memory 11b up to then to the state of selecting the output of the line memory 11a. As a result, the image data D23 is interpolated by the image data D22 (see an upward arrow). At that time, the image data D23 is interpolated in such a manner that the ineffective signal is interpolated by the image data D22 of the boundary pixel inside the effective pixel region because the image data D22 corresponds to the boundary of effective lines.
Then, generating the contour emphasizing signal S2 based on the coefficients as k2=−1, k5=2 and k8=−1, the high-frequency component is not detected because a signal of the pixel P21 is equivalent to the pixel P22 at the timing t21. Further, the high-frequency component is not detected because a signal of the pixel P23 is equivalent to the pixel P22 at the timing t24. In other words, there is no generation of the abnormal contour emphasizing component in the contour emphasizing signal S2. In the contour-emphasized image signal image data), a region represented by SA2 can be used as the effective signal.
As described so far, when the contour emphasizing process in the horizontal scanning direction shown in
In the embodiment 2, the input/output switching circuit 15 is not provided, and an image boundary detecting circuit 14 is adapted to control a filter coefficient of a contour component signal generating circuit 13a. The contour component signal generating circuit 13a is adapted to execute the operation of the contour emphasis in such a state that the filter coefficient with respect to the line on the upper side of the line of attention is set to “0” when the image boundary detecting circuit 14 detects an upper-side image boundary in the vertical direction, the filter coefficient with respect to the line on the lower side of the line of attention is set to “0” when the image boundary detecting circuit 14 detects a lower-side image boundary, the filter coefficient with respect to the pixel on the left side of the pixel of attention is set to “0” when the image boundary detecting circuit 14 detects a left-side image boundary in the horizontal direction, and the filter coefficient with respect to the pixel on the right side of the pixel of attention is set to “0” when the image boundary detecting circuit 14 detects a right-side image boundary. Any other structural element in the present embodiment corresponds to those in the constitution shown in
The timing charts of
At the timing t11 shown in
At the timing t14, the image boundary detecting circuit 14 detects the right-corner pixel, and the contour component signal generating circuit 13a sets the array of the coefficients as k4=−1, k5=1 and k6=0, while setting any other coefficient to “0” based on the detection result. As a result, the image data D13 is interpolated by the image data D12 (see the upward arrow). At that time, the image data D13 is interpolated in such a manner that the ineffective signal is interpolated by the image data D12 of the boundary pixel inside the effective pixel region because the image data D12 corresponds to the boundary of the effective pixels.
At the timing t21 in
At the timing t24, the image boundary detecting circuit 14 detects the lower-corner line, and the contour component signal generating circuit 13a sets the array of the coefficients as k2=−1, k5=1 and k8=0, while setting any other coefficient to “0” based on the detection result. As a result, the image data D23 is interpolated by the image data D22 (see the upward arrow). At that time, the image data D23 is interpolated in such a manner that the ineffective signal is interpolated by the image data D22 of the boundary pixel inside the effective pixel region because the image data D22 corresponds to the boundary of the effective lines.
According to the embodiment 2, the array of the coefficients is switched over depending on whether or not the pixel of attention corresponds to the boundary. Therefore, an effect similar to that of the embodiment 1 can be exerted without providing the input/output switching circuit 15 for switching the selectors.
The embodiments 1 and 2 recite the method of emphasizing the contour without reducing the effective pixels provided that the input image size is by n pixels or (n+1) pixels in an x direction and m pixels or (m+1) pixels in a y direction when a standard output image size is by n pixels in the x direction and m pixels in the y direction.
In contrast to that, the conventional method does not cause any particular problem in the contour emphasis as far as the input image size is by (n+2) pixels or more in the x direction and (m+2) pixels or more in the y direction relative to the standard output image size, on which an embodiment 3 of the present invention is premised.
When the image size adjusting signal Sc indicates the increase of the number of the pixels by two or increase of the number of the lines by two relative to the standard, the action of the image boundary detecting circuit 14 is halted. More specifically, provided that the input image size is by n pixels or (n+1) pixels in the x direction and m pixels or (m+1) pixels in the y direction in the case of the standard output image size having n pixels in the x direction and m pixels in the y direction, the image boundary detecting circuit 14 is activated. On the contrary, the image boundary detecting circuit 14 halts its action in the case of the input image size is by (n+2) pixels or more in the x direction and (m+2) pixels or more in the y direction.
According to the embodiment 3, the effective pixel region is set to be larger than the output image size so as to emphasize the contour based on the actual effective data. In such a manner, the contour can be naturally emphasized without reducing the output image size. In other words, the contour emphasizing signal can be constantly generated in an optimum state.
In the constitution shown in
The basic embodiments were thus far described. In addition, the present invention can be implemented in the following modes.
1) In the described embodiments, the filter region has 3×3 pixels centered around the pixel of attention, in place of which the filter region may have 5×5 pixels. In that case, the same process is executed to two pixels each in the right, left, upper and lower directions of the image boundary.
2) In the case of the filter region having 5×5 pixels, the image size adjusting signal Sc is activated when the input image size is by (n+4) pixels or more in the x direction and (m+4) pixels or more in the y direction in the embodiment 3.
3) The embodiments can be applied to such a case that a filter region having an optional size of nx pixels×ny pixels is provided and an array of coefficients corresponding thereto are used.
4) The structural elements may be respectively realized by hardware or software, or may be realized in a hybrid manner using both the hardware and software.
While there has been described and illustrated in detail, it is to be clearly understood that this is intended be way of illustration and example only and is not to be taken by way of limitation, the spirit and scope of the present invention being limited only be the terms of the following claims.
Number | Date | Country | Kind |
---|---|---|---|
P2004-154501 | May 2004 | JP | national |