JPS6141464B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a high pixel density conversion device for binary image information on the receiving side of an image transmission device such as a facsimile.

In this type of image transmission device, the scanning line density is
As the sampling density decreases from 8 lines/mm to 6 lines/mm to 4 lines/mm, the original image is sampled at a lower sampling density, and the number of image signals to be transmitted decreases. In order to improve the efficiency of the transmission path, we would like to utilize encoding technology that utilizes image correlation and use as low a scanning line density as possible. When converted into an image, it becomes as shown in FIG. 3, and the unevenness of the border is more noticeable than in the original image.
Here, one square in Figure 3 corresponds to one sampled pixel, and its size is 1/4 in both width and height.
mm.

Therefore, the low pixel density image signal obtained by scanning the original image at a low scanning line density of 4 lines/mm on the transmitting side is converted to a high pixel density image, for example, 4 times as high, on the receiving side to eliminate the unevenness of the border. I would like to reduce the number of pixels and obtain an image with the same quality as the high pixel density image that has been scanned and transmitted at a high scanning line density of 8 lines/mm on the transmitting side.

To convert to 4x pixel density, convert 1 pixel to 2x
If the original pixel is a black pixel, all 2 x 2 pixels are converted into black pixels, and if the original pixel is white, all pixels are converted into white pixels.The conventional method, for example, as shown in Figure 3, However, even if the high pixel density conversion is performed, the image quality is not improved in any way. In Figure 4, large pixels separated by thick lines are 1/4 mm x 1/4 mm low pixel density pixels, and pixels separated by thin lines are 1/8 mm x 1/8 mm high pixel density pixels and large pixels. Four of the small pixels are 2×2 pixels. The same applies to FIGS. 5 and 6, which will be described later.

In the present invention, after each pixel is divided into small pixels for a binary image signal, the small black pixels that are connected based on the logic determined by the black and white arrangement of surrounding pixels are unconnected. A thinning process is performed to convert small white pixels into small white pixels within a range where the small white pixels do not have to be This is a high pixel density conversion device that performs thick processing to convert pixels into small and black pixels within a range where pixels are not combined.

In an image expressed by large pixels with a low pixel density, the smooth boundary in the original image is represented by a step-like uneven boundary, but according to the present invention, by processing the fineness, it is possible to Some of the small black pixels are converted to small white pixels, which are then returned to the original thickness by thickening processing, resulting in a high pixel density image with smooth boundaries similar to the original image.

FIG. 1 is a block diagram when the present invention is applied to the receiving side of a facsimile transmission device.

The transmitted encoded image signal is stored in register 1.
Once stored in
Then, the thick processing circuit 8 is used to smooth the step-like unevenness of the boundary, and the recording conversion circuit 5 is used to obtain an output image 6.

The control circuit 7 generates a clock, a synchronization signal, and a control signal to control each circuit.

The high pixel density conversion circuit shown in FIG. 13 equally divides one pixel into four small pixels as shown in FIG.

Before entering into an explanation of the operations of the thinning processing circuit and the thickening processing circuit, the thinning processing and the thickening processing will be explained using specific examples.

FIGS. 7, 8, and 9 show examples of narrowing and thickening simple patterns. The pattern a on the left is the original pattern, and the black pixels adjacent to the pixel are converted to white pixels to form the narrow pattern b in the center. Among the white pixels of the narrow pattern in the center, those adjacent to black pixels are converted to black pixels and become the thick pattern c on the right side. However, as seen in the right vertical line in Figure 7b, in the narrowing process,
A black pixel whose pattern connection would be broken if converted to a white pixel is not converted to a white pixel. Furthermore, in the thickening process, white pixels, which would result in unconnected black pixels joining together when converted to black pixels, as shown by the left vertical line in FIG. 7c, are not converted to black pixels.

As seen in FIG. 9, the thinning process and the thickening process have the characteristic of smoothing the stepped boundaries of the pattern.

FIG. 10 shows an example of a narrowing processing circuit. This circuit converts black pixels at the boundary into white pixels within a range where the connectivity of the image is preserved.

In FIG. 10, the output of the high pixel density conversion circuit 3 in FIG. Skeleton ROM10
If it is the final point, a signal k is output. Here, n is the number of pixels in one line, and the final point is a point that cannot be converted to a white pixel, that is, a point that cannot be deleted because the connectivity of the image will not be preserved if the point is converted to a white pixel. When the logical product i.d.f=1, since d=0 and f=0, both the left and upper sides of black pixel i are black pixels, and i is not a boundary point between the left and upper sides. Also, when the logical product k・i=1, k
=0, therefore black pixel i is the final point.

The outputs i, d, f+k, and i of the OR circuit 15 are the left and upper boundary points, with those not at the final point removed, and are input to the 2n+3 bit shift register 12. Similarly, shift register 1
2. Skeleton ROM13, AND circuit 17,1
9. Using the OR circuit 18, delete the right and lower boundary points that are not the final point. However, this final point includes the final point j'=k·i determined by the skeleton ROM 10 and stored in the n+2 bit shift register 11, and the newly generated final point k′·i after deleting the left and upper boundary points. There are two ways of i′. Output of OR circuit 18 i′・b′・h′+k′・
i'+j' is the result of converting all black pixels on the boundary except the final point to white pixels.

FIG. 11 shows an example of the skeleton ROMs 10 and 13. Pixel a is in the higher address 4 bits,
b, c, d, pixel e in the lower address 4 bits,
Make f, g, and h correspond. The x mark in the address section is
Indicates a bit that can be either 0 or 1. An o mark at the intersection means that the data k at that address is 0, and no mark means 1.
.

With this narrowing circuit, the uneven image shown in FIG. 4 becomes the smooth image shown in FIG. 5.

FIG. 12 shows an example of a thick circuit. This is before and after the thinning circuit 21 which has already been explained in detail.
It is equipped with NOT circuits 20 and 22. The thickening process can be accomplished by inverting the black and white of the image, performing narrowing processing, and then reversing the black and white. With this thickening circuit, the image in FIG. 5 is thickened to approach the size of the original image in FIG. 2.

In the above explanation, the configuration and operation of the narrowing processing circuit have been explained using the example shown in FIG. 10, but the narrowing processing circuit that can be used in the present invention is not limited to this example. , various thinning processing methods shown in IEICE Pattern Recognition and Learning Study Group material PRL75-66 "Various considerations on thinning methods" by Tamura, etc.
Of course, various known methods can be used.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of a facsimile receiving device using the high pixel density conversion device of the present invention, Fig. 2 is an example of an original image, Fig. 3 is a sample of the original image, and Fig. 4 is a sample of the original image. The figure shows a high pixel density conversion using the conventional method, Fig. 5 shows a narrowed version of Fig. 4, Fig. 6 shows a high pixel density conversion of Fig. 3 using the present invention, and Fig. 7. Figures 8 and 9 are examples of thinning and thickening processing, Figure 10 is an example of a narrowing circuit, and Figure 11 is a skeleton used for it.
A configuration diagram of the ROM, FIG. 12 is an example of a thick circuit, and FIG. 13 is a diagram showing the positional relationship between low pixel density pixels and high pixel density pixels. In the figure, 1... register, 2... decoding circuit, 3... high pixel density conversion circuit, 4, 21... thinning circuit, 5... recording conversion circuit, 6... recorded image, 7... control circuit, 9,1
2...2n+3 bit shift register, 10,1
3...Skeleton ROM, 11...n+2 bit shift register, 20, 22...NOT circuit, 1
4, 16, 17, 19...AND circuit, 15, 1
8...OR circuit, 8...thicker circuit.

Claims (1)

[Claims] 1. High pixel density conversion means for dividing each pixel into small pixels for a binary image signal consisting of a first and second level, assuming that the first level is white and the second level is black. and, for the divided image signal, convert small black pixels into small white pixels within a range in which connected small black pixels do not become unconnected based on logic determined by the black and white arrangement of surrounding pixels. a processing means, for the narrowed image signal, converting small white pixels into small black pixels within a range in which unconnected small black pixels are not combined based on logic determined by the black and white arrangement of surrounding pixels; 1. A high pixel density conversion device comprising: thick processing means.