JPH05260314A - Electronic storage processing method for photographic image - Google Patents

Abstract

PURPOSE:To store the images with high picture quality and at the rational cost by making great account of the low band components when one to ten million picture elements re coded in a standard coding system and setting the coding compression ratio at 1/8-1/80. CONSTITUTION:A DCT part 102 applies the discrete cosine transformation to a given image part every (8X8) matrices. At this time, the matrices are arranged so that the horizontal space frequency is set at a low and high levels at the left and right sides respectively and that the vertical space frequency is set at low and high levels at the upper and lower sides respectively. Then the numerical value set at the point in the vicinity of the upper left corner represents the low bands. A quantizing part 103 quantizes the output of the part 102 based on a quantization table 104. When the relationship is tested among the compression ratio, the number of picture elements and the picture quality, the substantially satisfactory picture quality is secured with a million or more number of picture elements and the compression rate of 1/64. Then one to ten million picture elements and the compression ratio 1/8-1/80 are defined best in regard of the memory cost.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for electronically storing a photographic image for electronically storing the photographic image.

[0002]

2. Description of the Related Art Conventionally, as an example of digitizing a photographic image and storing it after encoding, there is an electronic still camera that stores the image in an LSI memory. The number of pixels is 300,000-400,000 (the total number of R, G, B pixels), and the compression rate by encoding is 1/2 or more and 1/8.
It is less than. 5 sheets of high-definition mode to about 21 sheets of high-compression mode can be stored in a 1 MB LSI memory, which is partially used.

[0003]

However, since the number of pixels is small in the conventional case, the image quality is lower than that of general silver halide photography, and it has not been sufficiently spread. In addition, although the silver salt film is also stored without being encoded, it requires a large amount of memory, so that it is economically burdensome and has not spread.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method of digitizing an image of a quality satisfactory level within an economically acceptable range and storing the digitized photographic image.

[0005]

In the present invention, a coding method which is being standardized by, for example, the Joint of Photographic Experts Group (JPEG) is used, and the compression ratio and the number of pixels by the coding method are In consideration of the condition of 1 and the parameter value of the quantization matrix at the time of encoding, the number of pixels is 1 to 10 million, and when encoding by the standardized encoding method, the low frequency component is emphasized and encoded, and the compression rate is 1 / 8 to 1/80.

[0006]

The present invention is characterized in that the number of pixels is set to 1,000,000 to 10,000,000, the low-frequency component is emphasized in encoding, and the compression rate by encoding is set to 1/8 to 1/80. It differs from the prior art in the number of pixels and the compression rate.

[0007]

1 is an example of an apparatus to which the present invention is applied, 1 is a storage medium for accumulating image information from a film or the like, 2 is a drive for reading / writing information from the storage medium, and 3 is a display device. Is a display interface for performing interface processing with 4), 5 is a bus, 5 is a controller for controlling circuits on the bus, 6 is a codec, and 7 is a communication interface for transmitting and receiving via a network.

In order to operate this, the storage medium 1 in which the digitized image information is written is put in the drive 2 capable of reading and writing information from the storage medium 1, and the display method is instructed. According to this instruction, the image information in the storage medium of the drive 2 is selected and displayed on the display device. Here, as the drive 2, various storage media such as magneto-optical and magnetic can be used. For example, if it is a 3.5-inch optical disk, then 1
With a capacity of 28 megabytes, it is highly suitable for small systems.

The image information is created by writing information obtained by converting a silver salt film into an electron by photoelectric conversion, electronic still camera information, frame information from a video camera, image information created by a computer, etc. in a storage medium. Since the image information has a very large amount of information, it is encoded to compress the information amount in order to use the storage medium efficiently. Therefore, at the time of display, the coded image information from the storage medium is decoded by the codec 6 into the displayable image information.

The image information has a large amount of data, which increases the cost of the memory. The higher the image quality required, the larger the number of pixels required, and the memory cost increases. On the other hand, the JPEG (Joint of Photographic Ex
perts Group) is being standardized (scheduled for 1992).

FIG. 10 shows a block diagram of the JPEG baseline system. Reference numeral 101 in the drawing is image data obtained by digitizing an image, and for example, 8-bit information is given to each pixel for each color component of yellow, magenta, cyan, and black. 102 is a discrete cosine transform (Disc
rete cosine transform) section, 103 is a quantization section, and 10
Reference numeral 4 represents a quantization table.

In the DCT unit 102, when the discrete cosine transform is performed for each 8 × 8 matrix for a given image portion, the horizontal spatial frequency has 8 components and the vertical spatial frequency also has the same. Eight components can be arranged in a matrix. At this time, the horizontal spatial frequency is arranged so that the left side is low and the right side is high, and the vertical spatial frequency is a matrix arranged so that the upper side is lower and the lower side is higher. The numerical value is a value representative of the low frequency component referred to in the present application.

In the quantizing unit 103, the DCT unit 10
2 is associated with the quantization table 104 in the form of an 8 × 8 matrix in a form corresponding to the output from the DCT unit 10
The value of each position in the matrix in the output from 2 is compared with the value of the corresponding position in the matrix in the quantization table, and the former value is quantized. More specifically,
If the value at a certain (i, j) position on the matrix, which is the output from the DCT unit 102, is "100", the corresponding value at the (i, j) position on the quantization table 104 is " 25 ", the quantization result is" 4 ".
(That is, 100/25). In this case, the value from “88” to “112” is “4”. If the value at a certain (i, j) position in the quantization table is large, the compression rate will be large, but the error in reproduction will be large and the quality will deteriorate (the value in the quantization table is "50"). Is "75"
The quantization result from "1" to "124" is "2", and if the value in the quantization table is "5", "98" to "10"
The quantization result up to 2 "is" 20 "). Needless to say, selecting a small value in the vicinity of the upper left corner in the quantization table means that quantization is performed with emphasis on low-frequency components (horizontal spatial frequency and vertical spatial frequency). means.

Reference numeral 105 in FIG. 10 indicates the obtained DC
Coefficient, 106 is the inter-block prediction unit, 107 is the obtained A
C coefficient, 108 represents a zigzag scanning unit. The zigzag scanning unit 108 zigzags the 8 × 8 quantized result shown in the lower left of FIG.

Reference numeral 109 is a Huffman code table, 110 is a Huffman coding unit, 111 is a Huffman coding unit, 112 is a byte stuffing unit, 113 is a marker code, signaling parameters, and 114 is a combining unit.

When storing an image, the conditions of the compression rate and the number of pixels by this encoding method and the parameter value of the quantization matrix at the time of encoding are very important. The matrix shown in FIG. 8 is usually used as the quantization matrix value. FIG. 9 shows an example of a quantization matrix that emphasizes low frequencies. Using each quantization matrix,
When comparing the image quality when the image is compressed to 1/50, it is more
It can be seen that the image quality is far superior without the occurrence of false contours (contours that do not actually exist).

Under the above conditions, an opinion test was conducted by 10 image quality experts to examine the relationship between the compression rate, the number of pixels, and the image quality. The results are shown in Figure 2. It can be seen that almost satisfactory quality can be obtained when the number of pixels is 1,000,000 or more and the compression rate is 1/64 or less. Furthermore, considering the memory cost,
As shown in FIG. 3, it is concluded that the optimum number of pixels is 1 to 10 million and the compression rate is 1/8 to 1/80. Number of pixels 400
Even from the result of examining the image with the compression ratio of 1/16, it is found that this level is satisfied in most cases.

Regarding the image retrieval method, the album 21-i displayed on the screen as shown in FIG. 4 is selected by the mouse, and then the image information 22-i of the album is displayed in FIG.
Is displayed, and when one or several images are selected, the selected image information is displayed. The albums displayed on the screen are usually sorted in the order of year, date, and time, but they can also be sorted by genre such as flowers, mountains, and families. It is also a system that supports the digitization of photographic images. The electronic version of the film is shown in FIG. 6 and the print station is shown in FIG. The digitization of film is
The film 23 is developed (24), digitized by a photoelectric conversion element (25), and written in a storage medium 26 together with date information and the like. Electronic information 27 such as an electronic still camera can also be accepted. Also, if the film is put out at the developing / electronicizing station, it is possible to provide a service for transmitting the film to an electronic album at home through a telephone line or ISDN. In the print station shown in FIG. 7, the storage medium 26 in which the print information is written is received, and the printing device 28 makes a hard copy through the editing processing unit 27. Also,
The print station can also provide a service of receiving image information and its print information through a telephone line or ISDN and outputting a hard copy.

[0019]

As described above, according to the present invention,
In the method of digitizing and storing a photographic image, the number of pixels is 100.
10 to 10 million, encoding with emphasis on low frequency components and compression rate 1
Since it is set to / 8 to 1/80, there is an advantage that images can be stored with high quality and at a reasonable cost.

[Brief description of drawings]

FIG. 1 shows a schematic configuration diagram of an example of an apparatus to which the present invention is applied.

FIG. 2 shows an opinion test result of the relationship between the image quality, the number of pixels, and the compression rate.

FIG. 3 shows the relationship between the image quality, the number of pixels, and the compression ratio in consideration of cost.

FIG. 4 shows an example of an initial screen of an electronic album.

FIG. 5 shows a screen example after selecting an album from the initial screen of the electronic album.

FIG. 6 shows an explanatory diagram of a film developing / electronicizing station.

FIG. 7 shows an explanatory diagram of a print station.

FIG. 8 shows a conventional quantization matrix.

FIG. 9 shows an example of a low-frequency-oriented quantization matrix.

FIG. 10 shows a block diagram of a JPEG baseline system.

[Explanation of symbols]

 1 Storage Medium 2 Drive 3 Display Interface 4 Bus 5 Controller 6 Codec 7 Communication Interface

Claims (1)

[Claims] 1. A method of digitizing and storing a photographic image, wherein the number of pixels is 1 to 10 million, and when encoding by a standardized encoding method, low-frequency components are emphasized and the compression rate by encoding is 1. / 8 to 1/80. A method of digitizing and storing a photographic image.