JP3244406B2 - Image processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

TECHNICAL FIELD The present invention relates to a color image,
The present invention relates to an image processing apparatus that binarizes each color component such as R, G, B or cyan, magenta, and yellow. In particular, R
The present invention relates to an image processing apparatus that binarizes each color such as a GB signal by a dither method.

[0002]

2. Description of the Related Art In recent years, with the personalization of office equipment such as personal computers and word processors, there has been an increasing demand for capturing and processing color images with a CCD camera or the like. However, when an image is captured using a CCD camera or the like, unlike a facsimile or a scanner, the light source is not constant, so that it is necessary to adjust the brightness of the image and the contrast.

On the other hand, images captured by a camera are RGB
Three primary colors or their complementary colors, cyan, magenta,
In yellow (hereinafter referred to as CMY), each is binarized and displayed or printed. At this time, in order to represent an image as close as possible to the original image, it is sufficient that pseudo gradation can be expressed even when binarizing.

As the pseudo gradation expression method, a dither method and an error diffusion method are well known. In the dither method,
In a predetermined set of several pixels, the threshold value is changed for each pixel according to a certain pattern. Thereby, the number of dots exceeding the threshold value in the set changes according to the pixel value, so that it looks as if there is a gradation.

[0005] By the way, when binarization is performed by such pseudo halftone processing, there is a disadvantage that the outline portion is blurred. For this reason, as disclosed in Japanese Patent Application Laid-Open No. 3-145273 (H04N1 / 40), it is necessary to enhance the contour to improve this. This contour enhancement is R
When applied to each of GB, there is a problem that the circuit scale is increased, and there is a problem that hue shift occurs at an outline portion. Therefore, as shown in FIG. 2, a color natural image is separated into a luminance signal and a color difference signal, Only these processes are performed.

In FIG. 2, reference numeral 1 separates a luminance signal and a color difference signal from a color TV signal which is an NTSC signal. Reference numeral 2 denotes an outline emphasizing circuit for emphasizing the outline of the luminance signal. Reference numeral 9 denotes a contrast adjustment circuit for adjusting the contrast of the luminance signal. Reference numeral 3 denotes a conversion circuit that converts a luminance signal and a color difference signal into RGB signals.

[0007] 5 uses the dither method to convert the RGB signals into
This is a circuit for converting a RGB signal into a binary pseudo-halftone signal.

[0008]

In the image processing apparatus shown in FIG. 2, when contour enhancement and contrast adjustment are performed using a luminance signal, and then converted into RGB and then binarized by a dither method, an unnatural image such as a whitish background may occur. A dot may appear. An example is shown in FIG.

FIG. 3A shows a state in which an image of a person is captured by a CCD camera or the like, processed according to the conventional system shown in FIG. 2, and the contrast adjustment is completed. Since the person (62) appears darker than the background (61), as a result of increasing the contrast, the background is completely white. FIG. 3B shows the result of conversion from the luminance and color difference signals to RGB and binarization.

As described above, an unnatural point (63) appears on the background that should be white. This is because the upper limit value that can be taken during the contrast adjustment (25 for 8 bits)
This is because when the data saturated beyond 5) is converted from the luminance and color difference signals to RGB, the conversion result (RGB) becomes slightly smaller than the upper limit value due to the conversion error.

This will be described with reference to FIG. The data handled here are RGB, luminance, and color difference signals (Cr, Cb).
It is assumed that both are 8 bits. That is, each possible value is 0 to 255 (the color difference signal is -127 to 128). Here, it is assumed that there is a pixel having a value as shown in FIG. This pixel has a luminance value of 150 only, and the chrominance signal is 0 for simplicity.

Generally, the contrast of an image captured by a CCD camera is low. Here, it is assumed that the luminance is doubled to increase the contrast. Then, as shown in FIG.
The brightness value exceeds the upper limit value 255. Since a value exceeding 255 cannot be handled, the luminance of the pixel is set to 255 (that is, white).

After this,

[0014]

(Equation 1)

The data is converted into RGB according to the equation (1). If this conversion is ideal, the upper limit of each of RGB becomes 255 as shown in FIG. In this case, there is no problem because the saturated value of the luminance signal has the same value (255) after RGB conversion as when each of RGB is saturated.

However, when the conversion matrix of the formula (1) for converting the luminance and chrominance signals into RGB is realized by hardware, the circuit scale is limited (there is also a cost limitation).
It is difficult to create an ideal matrix from, and an error occurs in the conversion. Due to this conversion error, the value after conversion is
As shown in FIG. 4 (e), the value exceeds 255 or is insufficient.

If each color is binarized by the dither method, one of the threshold values of the dither happens to be located at the small difference. This part appears as a dot. That is, if the threshold is higher than the GB signal, this portion becomes a red point. As described above, as a result of the contrast adjustment, a colored point appears in a portion that should be saturated and stopped and become pure white.

The point generated by this conversion error is called conversion error noise. In this example, the contrast is adjusted by multiplying the luminance and RGB components of the pixel by a constant, but increasing the contrast increases the difference between the maximum value and the minimum value in each pixel. Even when other methods are used, the contrast adjustment has a property that each pixel value is easily saturated.

When the contrast adjustment is performed, since a specific portion (for example, a human face) in the image is used as a reference, the other portions (such as the background) tend to be saturated, and conversion error noise often appears.

[0020]

The above problem is caused by the fact that the contrast adjustment, in which the pixel value is likely to be saturated, is performed before the conversion of the luminance and color difference signals into RGB. In order to solve this, the contrast adjustment is performed after the conversion into RGB. [Operation] According to the present invention, the occurrence of the conversion error noise described above can be prevented.

[0021]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The configuration of the present invention will be described with reference to the first embodiment of FIG. Here, conditions such as the number of bits for each color component are the same as in the conventional description. The input color image is separated into a luminance signal and a chrominance signal by a luminance / chrominance separation circuit (1).

The outline of the separated luminance signal is emphasized by an outline emphasizing circuit (2). This compensates for the blurring of the outline in the dither processing while preventing the circuit scale from expanding. As a specific example of the edge enhancement, there is a method of calculating the Laplacian of the luminance in the original image and adding it to the original image. Next, an RGB conversion circuit (3) converts the luminance signal and the color difference signal after contour enhancement into RGB signals.

This is to finally obtain binary data of RGB. The state of the converted value at this time is shown in FIG.
(A). The difference due to slight error due to conversion is RG
B appears in each color component. Next, the contrast adjustment circuit (4) performs contrast adjustment for each of the RGB colors.

The contrast adjustment will be described with reference to FIG. Here, it is assumed that the value of RGB in each pixel is doubled in order to increase the contrast. Then Figure 5
As shown in FIG. 5B, all the RGB components are saturated and exceed the upper limit, and the result after the contrast adjustment is as shown in FIG. 5C.

When comparing FIG. 4 (e) with FIG. 5 (c), the input data to the dither processing circuit in the conventional example of FIG. 2 is the same as FIG. 4 (d) when there is no error. Is input to the dither processing circuit (5). Thereafter, the dither processing circuit (5) performs binarization processing. At this time, the conversion error noise that has appeared in the related art does not appear.

As described above, the RGB signals are obtained from the luminance and color difference signals.
After the conversion, the adjustment of the contrast can prevent the occurrence of conversion error noise. A second embodiment of the present invention will be described with reference to FIG. In the second embodiment, after a color image signal is separated into a luminance signal and a color difference signal by a luminance / color difference separation circuit (1), a black level adjustment circuit (7) performs black level adjustment on the luminance signal.

Since the luminance signal in a color image captured by a CCD camera or the like rarely becomes 0 even at the smallest value, a constant value is subtracted from the luminance value of each pixel to adjust the black portion to luminance 0. This results in a tighter screen and better contrast. From the above equation (1),
Subtracting a constant value from the luminance signal is completely equivalent to subtracting the same value from each of the RGB components, and thus performing only with luminance is effective in reducing the circuit scale.

After the black level adjustment, processing is performed by the contour emphasizing circuit (2), and thereafter, the processing is the same as that of FIG. In the second embodiment, the black level adjusting circuit (7) is placed before the contour emphasizing circuit (2), but the order of processing may be reversed. A third embodiment of the present invention will be described with reference to FIG.

In the first embodiment, the luminance and color difference signals are
B signal, but in the third embodiment, RGB
It is not a signal but a CMY signal. That is, the YMC conversion circuit (8) is arranged before the contrast adjustment circuit (4). Generally, RGB is used for display on a display or the like, and Y or Y is used for output to a printer or the like.
M and C CMY signals are used.

If only printing is intended, the third embodiment is more efficient than the first embodiment shown in FIG. In the present embodiment, the luminance / color difference signal is directly converted into a CMY signal. However, the signal may be converted into an RGB signal once, subjected to processing such as contrast adjustment, and then converted into a CMY signal at the time of output.

Conversely, the signal may be converted into a CMY signal once and then converted into an RGB signal at the time of output. In this case, in addition to the configuration of FIG. 7, a conversion circuit from RGB signals to CMY signals or from CMY signals to RGB signals is required. In printing, three primary colors called CMY signals in this embodiment
(Black) may be added. In this embodiment, CMY
In the case of a signal, it also includes a CMYK signal.

[0032]

The present invention is intended to prevent a conversion error noise which is a specific problem that occurs in an image processing circuit which generates a conversion error at the time of conversion to an RGB signal or the like and which is binarized by a dither method. It is. In the present invention, the above-described conversion error noise can be prevented by a simple circuit change in which the contrast adjustment in which the pixel value easily saturates is performed after the conversion from the luminance signal and the color difference signal to each of the color components such as RGB. A high-quality image can be obtained.

[Brief description of the drawings]

FIG. 1 is a diagram showing a first embodiment of the present invention.

FIG. 2 is a configuration example of conventional image processing.

FIG. 3 is a diagram illustrating generation of conversion error noise.

FIG. 4 is a diagram illustrating a cause of generation of conversion error noise.

FIG. 5 is a diagram illustrating signal processing according to the present embodiment.

FIG. 6 is a diagram showing a second embodiment of the present invention.

FIG. 7 is a diagram showing a third embodiment of the present invention.

[Explanation of symbols]

(1): a luminance signal / color difference signal separation circuit; (2): an outline emphasis circuit; (3): a conversion circuit for converting a luminance signal and a color difference signal into RGB; ... contrast adjustment circuit, (5) ... dither processing circuit, (6) ... image captured with a person, (61) ... person, (62) ... background, (63) ... Conversion error noise, (7) black level adjustment circuit, (8) conversion circuit for converting luminance signals and color difference signals into CMY signals.

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-7-93563 (JP, A) JP-A-4-35360 (JP, A) JP-A-4-259165 (JP, A) JP-A-2- 210968 (JP, A) JP-A-9-65158 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N 1/40-1/409 H04N 1/46-1/64

Claims (2)

(57) [Claims] An image processing means for performing image processing on at least one of a luminance signal and a color difference signal;
Conversion means (3, 8) for converting the luminance signal including the image-processed signal and the color difference signal into signals of respective color components such as R, G, B or cyan, magenta, yellow and the like.
A contrast adjusting means (4) for performing contrast adjustment on the signal of each color component; and a pseudo halftone conversion means (5) for binarizing the signal of each color component subjected to the contrast adjustment using a dither method. An image processing apparatus comprising: 2. An image processing apparatus according to claim 1, wherein said image processing means performs an outline emphasis process on said luminance signal.