WO1996005692A1

WO1996005692A1 - Method of processing video bitonal images

Info

Publication number: WO1996005692A1
Application number: PCT/EP1995/003143
Authority: WO
Inventors: Marcello Niccolini
Original assignee: INFOTRONIC SpA
Current assignee: INFOTRONIC SpA
Priority date: 1994-08-09
Filing date: 1995-08-08
Publication date: 1996-02-22
Anticipated expiration: 1997-02-09
Also published as: ITMI941730A1; IT1274608B; ITMI941730A0; EP0792552A1

Abstract

The invention concerns a method for processing and storing bitonal video images, including digital scanned matrixes of white and black pixels belonging to a source image (IS), which has to be compressed in order to be transmitted and/or visualised with a reduced number of pixels. According to this method at least one matrix of the source image (IS), with preestablished sizes, is compressed in only one destination pixel with a level of grey corresponding to the ratio between the number of the black pixels and the number of the white pixels of the matrix. In this way the quality of the image is maintained and it is guaranteed its legibility.

Description

TITLE : METHOD OF PROCESSING VIDEO BITONAL IMAGES .

DESCRIPTION

Technical field The present invention refers to a method for processing bitonal video images.

In particular the invention refers to a method for processing and storing bitonal video images including scanned digital matrixes of black and white pixels belonging to an source image (IS) which has to be compressed in order to be transmitted and/or visualised with a reduced number of pixels.

The invention refers also to a device for the realization of the above mentioned method. Background art

As it is well known in the modern digital image storing or filing systems the documents are scanned with a certain number of so-called pixels. For bitonal images these pixels can be black or white. At present the resolution for document images reaches also 400 pixels for inch, indicated also with the term "dots per inch", or more shortly, dpi.

The sampled bitonal images must be successively stored in a file, from which it must be possible to retrieve and transmit them to a monitor or to trasmit them via facsimile.

There is successively the problem of reducing the memory space required to store and transmit these document images, without creating difficulties in visualising or processing them. The prior art techniques use some compressing methods established by a committee called CCITT (Comite Consultatif International Telegraphique et Telephonique) and registered as "Group 3" or "Group 4".

The CCITT is a committee which defines the various standards for the compression of bitonal documents usually utilized in the documents trasmissions via facsimile or in the images filing.

The method "Group 4" performs a better compression than the one of the "Group 3". In spite of that, in case of error in the compressed document the data after the error result to be unretrievable. The method of the "Group 4" is therefore used only with systems free from errors, for istance for the electronic filing of documents.

The method "Group 3" remains at present the most used one in the transmission via facsimile, since the overloaded thelephone line may introduce errors during the transmission of documents.

After their electronic storage, the documents composed of bitonal images should be visualized also on monitors, in particular on a computer monitor.

Most of the computer monitors used at present have resolutions lower than 100 dpi, even if they can visualise sixteen different colours. The resolution usually utilized for the storage of bitonal images is on the contrary of 300 dpi. It appears therefore necessary to effect a reduction of the sizes of the image to be visualised.

The scaling procedure is effected by a central processing unit CPU (Central Processing Unit) and requires a long time. For istance, in a personal computer equipped with a microprocessor Intel 80x86 such procedure may take even a second and this time is unacceptable for an high quality visualisation and efficient scanning. To reduce the time taken for image scaling, prior art techniques use different procedures to compress further the image to be processed, but these procedures reduce the image quality and often render it unreadable.

The technical problem on which the present invention is based relates to the finding of a method and a device with such characteristics as to allow to scale down a bitonal image and visualise this image on a monitor, in particulkar on a computer monitor, overcoming the limits and/or the disadvantages which are still present in the prior art techniques.

Summary of the invention

The solution idea at the basis of the present invention is the one of performing a processing method of bitonal images which maintains the spatial information of the image, creating an output image with several levels of grey, and which increases at the same time the reading level of the output image itself.

On the basis of such solution idea the technical problem is solved through a method as previously indicated and according to the characterizing portion of the annexed claim 1.

The problem is also solved with a deyice described in the enclosed claims 8 and following.

The characteristics and advantages of the method according to the invention are explained in the following description and in an example of realization with reference to a device illustrated in the enclosed drawings.

Brief description of the drawings

In these drawings : - Figure 1 shows schematically a gate architecture for the processing of bitonal images including a device according to the invention;

Figure 2 shows schematically the device for the realization of the method according to the invention;

- Figure 3 shows in detail the device of figure 2;

- Figure 4 shows an example of scaling down a bitonal image performed according to the invention;

- Figure 5 shows a flux diagram according to the method of the invention

With reference to such figures, with 1 it is globally and schematically shown an electronic device integrated on a semiconductor and working according to the method of the invention.

With particular reference to figure 2 the device 1 includes an image processor IIP which is also indicated with 2 and inserted between a source bus S and a destination bus D.

The source bus S conveys all the data referring to the source image IS, that is binary values, corresponding to the bitonal pixels (white/black) .

The destination bus ID collects all t,he data referring to the destination image ID, that is values included between 0 and 2*m+l corresponding to pixel at several levels of grey.

Advantageously, according to the invention, the source bus S and the destination bus D are DRAM buses (Dynamic Access Memory-bus) .

The image processor 2 acts as master of the source bus S when bits corresponding to the source image must be loaded, or it can act as master of the destination bus D when the scaled image rotated at a preestablished angle α must be written.

In figure 1 it is globally shown an architecture 10 of a bitonal images processing system including the device 1 with the image processor 2.

Such an architecture includes a memory DRAM 37 connected with a decompressor 38.

The memory 37 furnishes a stored document in compressed form to the decompressor 38, which provides to store it in a register 39 connected with the source bus S.

The destination bus D is connected with a processing unit CPU 40 and with a further register 41 of the VRAM type, which transmits the output data to a look-up table 42, that is to a read only memory.

The look-up table 42 provides successively to furnish to a monitor the correct output values, not represented being of a known type.

The image processor 2 includes a series of registers, described hereinafter, which must be programmed by the processing unit 40 before the scaling operation starts, according to the invention.

The method according to the invention allows to realize a scaling down of a digital image according to a factor n/m, where n and m are positive integers included preferably between 1 and 31 . The maximum scale down factor is 1/31, whilst the maximum scale up factor is 31/1.

The document scanned images in digital format have usually resolutions of 200 or 300 dpi, whereas the resolution of a monitor reaches only 120 dpi. It is therefore clear that the images need to be scaled down to be transmitted on a monitor .

Figure 4 shows a scaling down example with a resolution of a scanned image according to a scale factor equal to 1:3. This means that a matrix of three by three pixels of a source image IS is converted into one single pixel of a destination image ID.

According to the traditional techniques, if the initial matrix contains, as in the example of figure 4, a number of black pixels greater than the one of the white pixels, the destination pixel is made black and looses part of the space information of the source image.

On the contrary, according to the present invention, to the matrix of the source image corresponds advantageously a destination pixel with a level of grey corresponding to the ratio between the number of the black pixels and the number of the white pixels.

In this way the space information of the source image IS is maintained in the form of intensity of the level of grey in the destination image ID.

The method according to the invention works therefore similarly to the human eye, for istance when we look at a figure in a newspaper. In fact the photos published by the newspapers consist of many black and white dots very close together. The dots are so close together that the human eye cannot resolve them individually and perceives the figure as having several levels of grey.

Let us now see in greater detail the scaling down method according to the invention, starting from the horizontal scaling of the image.

With particular reference to the scheme of figure 5, the pixels of the source image are first multiplied by a number of times corresponding to the value of the numerator n of the scale factor and added together in scaled groups corresponding to the value of the denominator m of the scale factor.

An input I, for istance of m binary pixels (black/white) produces an intermediate output 01 of n pixels with values from 0 to m, that is with m+1 levels of grey.

For a scale factor 3:5 we obtain for istance the situation illustrated in the following table:

1 0 0 1 1

Input ( I ]

Multiplication by n=3 111 000 000 111 111

Adding together by m=5 11100 00001 11111

Intermediate Output (01) 3 1 5

The intermediate output 01 is successively vertically scaled adding together the corresponding lines in order to obtain a prestablished number of final output O lines . The final output 0 lines are obtained through the sum of the intermediate output 01 lines and are stored in a temporary line storage buffer.

In case the adding of the lines does not result uniform it is effected a duplication of the lines in groups with a total lower number of lines.

For istance, to obtain three final output 0 lines starting from five intermediate output 01 lines there are added and summed the lines 1-2, 4-5, together with the line 3 repeated twice, according to the following tables: In u .

Line 1

Line 2

Line 3

Line 4: 5 0 2

Line 5: 0 1 1

Output

Line 1: 3 4 7

Line 2: 8 10 0

Line 3: 5 1 3

The input binary m lines (white/black) are so changed in output n lines with pixels which have values from 0 to 2*m, that is with 2*m+l levels of grey.

In particular the values of the final output O pixels obtained with the above described method are written on the destination bus D in memory addresses* which depend on the rotation angle α and are sent to the look-up table 42 in order to select the monitor dots and form on it the destination image ID.

Once the internal registers have been programmed by the system CPU the scaling procedure can be performed automatically until the operation is completed. Once completed the interrupt signal is made active to inform the system CPU. Now there are described in detail the series of registers incorporated in the device 1 for the realization of the method according to the invention.

The image processor 2 includes a register 3 SBBA (Source Bitmap Base Address) which contains the start address of the source image IS bitmap.

The processor 2 includes also a register 4 SBP (Source Bitmap Pitch) which contains a "pitch" of the source image IS bitmap, and a register 5 SIW (Source Image Width) which contains the width of the source image IS bitmap.

The image processor 2 includes also different registers which contain the information concerning the destination image ID and which are listed here below.

A first register 6 DBBPP (Destination Bitmap Bits Per Pixel) contains the pixel depth of the destination image ID bitmap.

Preferably the sizes of the destination image bitmap may be 1,4,8 or 24 bpp (bit per pixel) .

A second register 7 DR (Destination Rotation) contains the rotation angle α of the destination image ID bitmap.

Advantageously such rotation angle α of the destination image bitmap may be either 0, 90, 180,* or 270 degrees.

A third register 8 DBBA (Destination Bitmap Base Address) contains the pixel address of the upper left corner of the destination image ID.

Other registers 9 DXF (Destination X Field) and 10 DYF (Destination Y Field) contain respectively the X and Y values concerning the horizontal and vertical sizes of the destination image ID bitmap. A sixth and seventh register 11 DW (Destination Width) and 12 DH (Destination Height) contain respectively the width and the height of the destination image ID.

Another register 13 DRO (Destination Raster Operation) contains the codes of sixteen different boolean operations to combine the final output data obtained with the method of the invention with the data already in the destination image ID bitmap.

A ninth register 14 HSF (Horizontal Scale Factor) and a tenth register 15 VSF (Vertical Scale Factor) contain the scale factors (that is the two numbers n and m) , which determine how the image is scaled horizontally and vertically.

Moreover a register 16 SABSA (Sealer Accumulation Buffer Start Address) contains the start address of the temporary storage space used by the processor 2 during the scaling operation.

The processor 2 can moreover contain a look-up table 17 including several scale values corresponding to the most common monitors in use.

To this series of registers corresponds a series of terminals for the external connection with the source bus S and destination bus D, in order to control the working of the image processor 2 during the scaling operations of a bitonal image.

A terminal 18 is connected with a 16-bit bidirectional data bus SData, e a further terminal 19 is connected with a 13- bit bidirectional address bus SAddr and both of them are connected to the source bus S.

A terminal 20 and further terminal 21 are connected with a couple of indicators SRas and SCas, which indicate the presence on the address bus SAddr of row addresses or column addresses.

A terminal 22 is connected with the read/write line SWE connected with source bus S, and a terminal 23 is connected with the data sampled line SOE connected with the source bus S.

A terminal 24 and a further terminal 25 are then connected with the lines SReq and SGnt, which indicate that an external master wants to become or may become master of the source bus S.

A terminal 26 is connected with the interrupt signal Int, which is made active whenever the image processor 2 finishes a scaling operation.

Similar and correspondent connections are designed between the terminals indicated with the number 27-34 and the destination bus D.

It must be pointed out that the bus DData connected with destination bus D has 32 bit.

Moreover further connections 33, 34 are designed for the lines DReq and DGnt, which have the task to indicate that the image processor 2 wants to become or may become master of destination bus D.

A further terminal 35 is connected with a line DOwn, which indicates when the image processor 2 is master of the destination bus D.

At last a terminal 36 is connected with a line DCS for chip selection, which is used by an external processor to read and write in the inside of the image processor 2 registers.

We can conclude that the method according to the invention and the corresponding device for its realization allow to obtain, starting from a source bitonal image IS scanned in a certain number of pixels, a destination image ID compressed according to a scale factor n:m and rotated by an angle α, maintaining the spatial information in the form of several levels of grey and increasing greatly the legibility of the image self.

Claims

1. Method for processing and storing bitonal video images including digital sampled matrixes of black and white pixels belonging to a source image (IS) , which has to be compressed in order to be transmitted and visualised with a reduced number of pixels, characterized by the fact that at least one matrix of the source image (IS) , with preestablished sizes, is compressed in only one destination pixel (ID) , having a level of grey corresponding to the ratio between the number of the black pixels and the number of the white pixels of the matrix.

2. Method according to claim 1, charaterized by the fact that the spatial information connected with a source image (IS) is maintained in form of intensity of level of grey in the destination image (ID) .

3. Method according to claim 1, characterized by the fact that the compression is realized according to a scale factor n/m and the pixels of the source image are firstly multiplied by a number of times corresponding to the value of the numerator n of the scale factor and added together in scaled groups corresponding to the value of the denominator m of the scale factor.

4. Method according to claim 3, characterized by the fact that the sizes in pixels of the source image matrix to be compressed are indicated by the denominator of the scale factor.

5. Method according to claim 3, characterized by the fact that an input signal (I) including m white and/or black pixels produces an intermediate output (01) of n pixels with different levels of grey.

6. Method according to claim 5, characterized by the fact that this intermediate output (01) is successively vertically scaled through the adding of the corresponding lines, in order to obtain a preestablished number of final output (0) lines.

7. Method according to claim 5, characterized by the fact that the above different levels of grey are m+1.

8. Electronic device for the compression of bitonal video images, including digital sampled matrixes of white and black pixels belonging to a source image (IS) , which has to be compressed in order to be trasmitted and/or visualised with a reduced number of pixels, characterized by the fact that it includes an image processor (2) inserted between a source bus (S) , which conveys the data referring to the source image, and a destination bus (D) , and that this processor (2) works on the matrixes of the source image (IS) to compress them in destination pixels (ID) having a level of grey corresponding to the ratio between the number of black pixels and the number of white pixels of the source matrixes.

9. Device according to claim 8, characterized by the fact that the source bus (S) and the destination bus (D) are of the DRAM type.

10. Device according to claim 8, characterized by the fact that the processor (2) includes several registers, of which at least one contains the value of the compression scale factor.

SUBSΗTUTESHEET(RULE26)