JPH0468772A - Picture data decoding device - Google Patents

Abstract

PURPOSE:To display lots of reduced patterns in less arithmetic quantity and time by extracting only a DC component of a data and multiplying a prescribed value from a ROM coefficient with an extracted data and interpolating the data. CONSTITUTION:A data is divided into a quantization matrix and a picture data by a division circuit 2 and the picture data subjected to fixed length processing is subjected to inverse quantization by an inverse quantizing device 4 by using a DC component 20 of a quantization matrix in a sent data. Then the data of the DC component subjected to inverse quantization is inversely converted. An address data is converted by using a coefficient ROM 5 and the converted data is interpolated according to the reduction rate of the displayed pattern and a required magnification depending on a picture element number M of one side of a data block subjected to orthogonal conversion. Thus, since only the data DC component is decoded, lots of reduced patterns are displayed in less calculation quantity and time.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention provides an image compressed using two-dimensional orthogonal transformation.
The present invention relates to an image data decoding device for displaying multiple images.

BACKGROUND OF THE INVENTION In recent years, there has been an increasing need to compress, transmit, and store image information, and technology has developed to display a large number of images obtained in this way simultaneously on, for example, a CRT and search for the desired image. being developed.

First, an encoding method for image compression using two-dimensional orthogonal transformation will be explained. FIG. 3 shows a block diagram of a general image data compression device. Digital image data is divided into two-dimensional plots by a block division circuit 8. The size of the blocks divided here is 8x8 pixels or 18x1
B pixels etc. are used. The divided data is subjected to two-dimensional orthogonal transformation by an orthogonal transformation circuit 9. Orthogonal transformations used here include discrete cosine transformation, KL transformation, and the like.

Since the data after orthogonal transformation has been transformed into the frequency domain, it is quantized by the quantizer 10 based on the visual characteristics of human spatial frequency sensitivity. The quantization performed here is performed by dividing the data after orthogonal transformation by a two-dimensional quantization matrix circuit 11. The quantized data is zigzag scanned (scanning the data at each point on the screen in the direction indicated by the arrow) in the order shown in FIG. 4 by the zigzag scan circuit 12, and converted into a one-dimensional data series. It will be done. The quantized data is variable length coded in the variable length encoder 13, and the first data to the last non-zero coefficient in the block is coded,
and a code indicating the end of the block (End Of Blo
ck) is added.

Next, a conventional decoder for decoding data encoded by the above encoding method will be described. FIG. 5 shows a block diagram of a conventional image data decoding device. Here, the data to be decoded consists of a quantization matrix portion representing quantization of data and a compressed image data portion, as shown in FIG. Further, the quantization matrix includes a DC component 20 and an AC component 21, and the image data portion also includes a DC component 22, an AC component 23, and an EOB 24 indicating the end of a block.

The data is divided by a division circuit 14 into a quantization matrix and image data. Since the compressed image data is variable-length encoded, the data is first converted to a fixed length by a fixed-length decoder. The decoding performed here is performed on entropy encoded data. For example, Huffman code, Fill code, etc. Since the fixed length data has been quantized, it is dequantized by the dequantizer 16. Here, the dequantizer 16 dequantizes the sent data using a quantization matrix contained in the data.

The dequantized data is structured into two-dimensional blocks such as 8×8 or 18×1B, as shown in FIG. 7, and one block consists of a DC component 25 and an AC component 26.

Here, if the inside of the block is not filled with decoded data, 0 is inserted. Data configured into two-dimensional blocks is subjected to two-dimensional inverse orthogonal transformation according to the size of the block by an inverse orthogonal transformation circuit 17. The inverse orthogonal transformed data is band-limited to less than half the sampling frequency of the image to be displayed by a low-pass filter 18 so that aliasing does not occur in the subsequent thinning process. The data band-limited by the low-pass filter 18 is thinned out by a thinning process 19 . Here, when displaying an image by reducing it by -N, the low-pass filter needs to be 2N to the sampling frequency of the image.
Bandwidth limitation is applied to - of N, and subsampling is performed to thin out to - of N.

Problems to be Solved by the Invention However, in the image data decoding device as described above, if a large number of reduced images are to be displayed on the screen, it is necessary to decode all the compressed data, which requires a large amount of calculation and calculation time. Requires.

SUMMARY OF THE INVENTION An object of the present invention is to provide an image data decoding device that displays a large number of reduced images with a small number of calculations.

Means for Solving the Problems In order to achieve the above object, the image data decoding device of the present invention specifies the address of the DC component in each block of digital image data compressed by orthogonal transformation and the DC component of the quantization matrix. a dividing circuit that separates the DC component in each block of digital image data at a designated address from the DC component of the quantization matrix; and a fixed length circuit that converts the extracted variable length data into a fixed length. A decoder and a quantization matrix D
A dequantizer for multiplying by the value of the C component, a coefficient ROM for dividing the dequantized DC component data by the number of pixels on one side of the block, and an interpolation process according to the reduction rate of the image to be displayed. The frame memory is configured to include an interpolator that performs the image processing, and a frame memory that stores display image data.

Effects In the above configuration, only the DC component of the compressed data is extracted and decoded, so the amount of calculation is extremely reduced, and a large number of reduced images can be displayed in a short time.

Embodiment Hereinafter, an image data decoding device according to an embodiment of the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram of an image data decoding device according to an embodiment of the present invention.

When displaying a reduced screen, the high frequency part of the AC component is not necessary. Furthermore, when attempting to decode the low frequency portion of the AC component, matrix calculations are performed during inverse orthogonal transformation, which increases the circuit scale. Therefore, in the present invention, only the DC component is used for decoding.

The stored data consists of a quantization matrix representing data quantization and a compressed image data portion, as shown in FIG. Further, the quantization matrix consists of a DC component 20 and an AC component 21, and the image data part also consists of a DC component 22, an AC component 23, and an EOB 24 indicating the end of a block. Therefore, the DC component 20 of the quantization matrix and the DC component 22 of the data
The address designation circuit 1 specifies an address for extracting only the data.

The data is divided into a quantization matrix and image data by a division circuit 2. Since the compressed image data has been variable length encoded, the fixed length decoder 3 first converts the data into a fixed length. The decoding performed here is performed on entropy encoded data. For example, Huffman code, Weyl code, etc. Since the fixed length data has been quantized, it is dequantized by the dequantizer 4. Here, the fixed length image data is dequantized by the DC component 20 of the quantization matrix in the sent data.

Next, the inversely quantized DC component data is inversely transformed. The two-dimensional inverse orthogonal transform is expressed as follows, where the encoded data matrix is [Y] and the decoded data matrix is [X].

[X co = [D] T [Y co [D co here,
[D is a matrix of MXM, and each element d of IJ. 1. For example, for the discrete cosine transform,
It will look like this:

Now, if the DC components are Va and a, then X n , a becomes X n , , = 1 / M X V @, 6. Therefore, the dequantized DC component data is M
Instead of processing the two-dimensional inverse orthogonal transformation of xM, if we divide by the number of pixels M on one side of the block shown in FIG.
An image compressed to 7M can be obtained. Here, the same value for all data (1 in the case of discrete cosine transform)
/M) is multiplied by all AC components by zero, address data conversion is performed using the coefficient ROM 5 without using a multiplier. Next, the converted data is interpolated according to the reduction ratio of the screen to be displayed and the necessary enlargement ratio determined by the number M of pixels on one side of the orthogonally transformed data shown in FIG. In other words, if the reduction ratio of the screen is 1/N and the number of pixels on one side of the orthogonally transformed data block is M, then the enlargement ratio is M/N.

For example, if the accumulated data is orthogonally transformed data in 8x8 blocks and the reduction ratio of the displayed screen is 1/4, the number of images per side will be doubled as shown in Figure 2. Interpolated so that

The interpolated data is written to the frame memory in screen units 27 according to the addresses shown in FIG. Once the frame memory is fully filled with multiple screens, it is read out for display.

As described above, according to the decoding device of the present invention, the data DC
Since only the components need be decoded, a large number of reduced screens can be displayed with a small amount of calculation and time.

Effects of the Invention As is clear from the above description of the embodiments, the decoding device of the present invention extracts only the DC component of the data without performing inverse orthogonal transformation when decoding an image signal, and converts the extracted data into a certain type using ROM coefficients. Since interpolation is performed by multiplying by a constant value, many reduced screens can be displayed with a small amount of calculation and time.

[Brief explanation of drawings]

Fig. 1 is a block diagram of an image data decoding device of the present invention, Fig. 2 is an explanatory diagram showing an example of interpolation processing, Fig. 3 is a block diagram of a conventional image data compression device, and Fig. 4 is a zigzag scan FIG. 5 is a block diagram of a conventional image data decoding device. FIG. 6 is an explanatory diagram showing the format of stored data. FIG. FIG. 8 is an explanatory diagram showing an example of thinning processing, and FIG. 9 is an explanatory diagram showing the address arrangement of screen data written to the frame memory. DESCRIPTION OF SYMBOLS 1... Address designation circuit, 2... Data division circuit, 3... Fixed length decoder, 4... Inverse quantizer, 5... Coefficient ROM1 B... Interpolation circuit,
7...Frame memory. Name of agent: Patent attorney Shigetaka Awano (1 person)

Claims (2)

[Claims] (1) An addressing circuit that specifies the DC component in each block of digital image data compressed by orthogonal transformation and the address of the DC component of the quantization matrix; A dividing circuit that separates the DC component from the DC component of the quantization matrix, a fixed length decoder that converts the extracted variable length data into a fixed length, and converts the decoded data into the value of the DC component of the quantization matrix. an inverse quantizer that multiplies by , a coefficient ROM that divides the inverse quantized DC component data by the number of pixels on one side of the block, and an interpolation circuit that performs interpolation processing according to the reduction rate of the image to be displayed. An image data decoding device comprising a frame memory in which display image data is stored. (2) The encoded DC component data is decoded by simply dividing it by the number of pixels on one side of the block, and then the interpolation process is performed in accordance with the reduction ratio of the screen to be displayed. The image data decoding device described above.