JPH05176171A - Method and device for compressing picture data - Google Patents

Abstract

PURPOSE:To realize picture data compression whose memory capacity is reduced in the system requiring stroboscopic photographing in a fixed length processing for picture data of a digital still camera based on the adaptive discrete cosine transformation. CONSTITUTION:This device is provided with an A/D converter 1, a DCT circuit 2, a quantization circuit 3, a buffer memory 4, a Huffman coding circuit 5, an inverse quantization circuit 6, and an encode quantity monitor circuit 7. Data read from the buffer memory 4 are inversely quantized at an inverse quantization circuit 6 by a same quantization table or coefficient as that when the data are quantized, the data are again quantized at a quantization circuit 3 with a new quantization table or coefficient and the result is written in the buffer memory 4 to fix the length of the code quantity.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data compression method and an image data compression apparatus for fixing a fixed length of image data of a digital still video camera (DSVC) by an adaptive discrete cosine transform (hereinafter referred to as ADCT). It is a thing.

[0002]

2. Description of the Related Art Two methods are currently considered for fixing a fixed length when compressing image data by ADCT, and each has the following characteristics. One is a real-time system, and FIG. 3 shows a block of the real-time system. In FIG. 3, 11 is an A / D converter and 12 is a DC
T circuit, 13 is a quantization circuit, 14 is a Huffman coding circuit, 15
Is a code amount monitor circuit, which converts an input video signal into a digital signal by the A / D converter 11 and compresses it in real time by the DCT circuit 12, and depending on the generated code amount, a quantization table or a quantization table. The coefficient to be multiplied by is changed and the quantization circuit 13, until the code amount reaches a specified value,
It is adjusted by the Huffman coding circuit 14 and the code amount monitor circuit 15. In this method, the input video signal needs to be constant until the fixed length is completed. The other is a frame memory system, and FIG. 4 shows a block of the frame memory system. In FIG. 4, 21 is A / D
Converter, 22 is a frame memory, 23 is a DCT circuit, 24
Is a quantization circuit, 25 is a Huffman coding circuit, and 26 is a code amount monitor circuit. The difference from the real-time method shown in FIG. 3 is that the input video signal is once stored in the frame memory 22 and is processed while being read out. That is, the video signal is once stored in the frame memory, the data is compressed by the DCT circuit 23 while reading it, and the quantization circuit is used.
24, the Huffman encoding circuit 25, and the code amount monitor circuit 25 perform fixed length conversion. According to this method, since the input video signal is once stored in the memory, only one screen of the video signal is required, and it is possible to deal with a flash photography or a fast-moving subject.

[0003]

However, in the above-mentioned real-time system, the input video signal must be constant until the fixed length is completed, and fixed length cannot be achieved in the case of photographing by stroboscopic light emission. There is a problem. Further, the frame memory system can be used for stroboscopic photography and a fast-moving subject, but the frame memory has a problem in that it is disadvantageous in terms of cost and size.
The present invention solves the above-mentioned conventional problems. In a system in which stroboscopic photography is indispensable, it is necessary to perform a fixed length by a frame memory method. It is an object of the present invention to provide an image data compression device.

[0004]

In order to achieve the above-mentioned object, the present invention provides (1) an image data compression method, wherein a digitized signal of an input video signal is quantized after discrete cosine transform and stored in a buffer memory. One screen is temporarily stored, the data read from the buffer memory is dequantized with the same quantization table (or the coefficient to be multiplied to the quantization table) as when it was quantized, and the data is stored in a new quantization table (or coefficient). ), It is quantized again and written in the buffer memory so that the code amount is fixed. (2) Further, the image data compression device includes an A / D converter that converts an input video signal into digital data and a discrete cosine conversion (hereinafter, referred to as D
A DCT circuit that performs CT), a quantization circuit that quantizes the DCT data, a buffer memory that can store the quantized data for one screen, and a Huffman that Huffman-encodes the data read from the buffer memory. An encoding circuit, a code amount monitor circuit for monitoring the Huffman-encoded data, an inverse quantization circuit for inversely quantizing the data read from the buffer memory, and an input to the quantization circuit for the DCT circuit. A switch for switching between the inverse quantization circuit and the inverse quantization circuit is provided.

[0005]

Therefore, according to the present invention, (1) the buffer memory for one screen is provided as compared with the real-time method, so that a fixed length can be achieved even for stroboscopic photography or a fast-moving subject. (2) Compared to the frame memory method, the memory is placed after the quantization circuit, so the memory capacity is small. Although the number of inverse quantization circuits is increasing, a Huffman decoder, an inverse quantization circuit, and an inverse DCT circuit are required in a camera capable of recording / reproducing, so that the inverse quantization circuit can be used in common.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows the structure of an image data compression apparatus according to an embodiment of the present invention. In FIG. 1, 1
A / converts the input video signal into digital data
D converter, 2 is a DCT circuit, 3 is a quantization circuit for quantizing DCT data, 4 is a buffer memory capable of storing the quantized data for one screen, 5 is a Huffman coding circuit, 6 is data Is a dequantization circuit for dequantizing, H is a code amount monitor circuit for monitoring Huffman-coded data, and 8 is a switch for switching.
The compression of the image data of the ADCT method is usually about 1 / n in the process of quantization when the input image data is compressed to 1/16.
4. In the process of Huffman coding, it is further compressed about 1/4, resulting in a total compression of 1/16. DCT and quantization are reversible and can be converted into original data. On the other hand, Huffman coding is lossy. FIG. 1 shows a buffer memory with a small capacity based on the above two principles.

Next, the operation of the above embodiment will be described. The input video signal is converted into a digital signal by the A / D converter 1, DCT and quantization are carried out in real time by the DCT circuit 2 and the quantizing circuit 3, and stored in the buffer memory 4 once. Since the amount of data at this time is after being quantized, it is about 1/4 of the original data. After that, the data in the buffer memory 4 is read out and encoded by the Huffman encoding circuit 5, and the generated data amount is monitored by the code amount monitor circuit 7 to obtain a quantization table (or coefficient) where the code amount becomes a specified value. Next, the contents of the buffer memory 4 are read out, dequantized by the dequantization circuit 5, requantized by a new quantization table (or coefficient), and written in the buffer memory 4. This is repeated until the code amount reaches a specified value. Here, rewriting of the buffer memory is performed by the procedure shown in FIG. First, the data in the buffer memory is as shown in FIG. First, the original data 1 is read and inversely quantized, and the requantized data is written in the empty area of the buffer memory. Then, the new data 2 is also written in the empty area in the same manner as shown in FIG. 2B. Next, since there is no empty area, the new data 3 is written on the original data 1. Similarly, when data 6 is completed, the data rewriting is completed as shown in FIG. 2 (c). This is repeated until fixed length is completed. Note that FIG. 2 is a schematic diagram, and the number of actual data is usually several M bytes.

[0008]

The present invention has the following effects as is apparent from the above embodiments. (1) Compared to the real-time method, it has a buffer memory for one screen, so fixed length can be achieved even for flash photography and fast-moving subjects. (2) Since it is placed in the buffer memory after the quantization circuit, the memory capacity is small. Although an inverse quantization circuit is required, a camera capable of recording / reproducing needs a Huffman decoder, an inverse quantization circuit, and an inverse DCT circuit, and therefore the inverse quantization circuit can also be used. Does not increase costs.

[Brief description of drawings]

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

FIG. 2 is a schematic diagram illustrating rewriting of a buffer memory according to an embodiment of the present invention.

FIG. 3 is a block diagram of a conventional real-time image data compression apparatus.

FIG. 4 is a block diagram of an image data compression device in a frame memory system in a conventional example.

[Explanation of symbols]

1, 11, 21, A / D converter, 2, 12, 23 ... Discrete cosine conversion circuit, 3, 13, 24 ... Quantization circuit, 4 ... Buffer memory, 5, 14, 25 ... Huffman coding circuit, 6 ... Dequantization circuit, 7, 15, 26 ... Code amount monitor circuit, 22 ... Frame memory.

Claims (2)

[Claims] 1. An adaptive discrete cosine transform (A
In a digital still video camera that performs data compression by the DCT method, the digitized video signal is quantized after discrete cosine transform (DCT) and one screen is temporarily stored in a buffer memory, and the data read from the buffer memory is quantized. Image data characterized by dequantizing with the same quantization table or coefficient as when it was done, requantizing the data with a new quantization table or coefficient, and writing it in the buffer memory to fix the code length. Compression method. 2. A digital still video camera for compressing data by an adaptive discrete cosine conversion method, an A / D converter for converting an input video signal into digital image data, and a discrete cosine for converting the digital image data into a discrete cosine. A transform circuit, a quantizer circuit for quantizing discrete cosine transformed data, a buffer memory capable of storing the quantized data for one screen, and a Huffman coding for coding the data read from the buffer memory. A circuit, a code amount monitor circuit for monitoring Huffman-encoded data, an inverse quantization circuit for inverse quantizing the data read from the buffer memory, and a switch for switching the input to the quantization circuit. Image data pressure characterized by having Compression device.