Research on Wavelet Transform Image Compression Mode(II)

(continued above)

肆, WAVELET image compression future development trend

First, strengthen its completeness in its structure.

Second, modify the program so that it can handle images with different mode ratios.

Third, support more colors. Can handle RGB colors, such as YIQ, HUV's color definition can be handled separately.

Fourth, to strengthen the ability to calculate so that it can support more video formats.

V. Use WAVELET conversion by eliminating high frequency data rate increase.

Six, increase a variety of WAVELET. Such as: discrete, zero-tree and so on.

7. Modify its mathematical encoder so that data can be converted between mathematical formulas and computer bits.

Eight, add 8X8 grid DCT mode, so that it can do JPEG compression.

Nine, increase the 8X8 grid DCT mode, so that they can overlap.

X. Increase trellis coding.

Eleven, increase zero tree.

Nowadays, the Academia Sinica has commissioned domestic academic units to study, and many research institutes have published their master's thesis. It is in full swing that foreign countries are conducting research. It is believed that the days of practical application are just around the corner.

Wu, the direction of image compression research

1. How the input device captures the real image and digitizes it.

2. How to convert digital image data into data types that are conducive to coding.

3. How to control the quality of decoded images.

4. How to choose the appropriate coding method.

5. The human visual system's response to images.

Wavelet analysis, whether it is a continuous wavelet transform as a mathematical theory, or a discrete wavelet transform as an analytical tool or method, still has many places that can be studied. It is a major breakthrough in tools and methods in recent years. Wavelet analysis is an important development of Fourier analysis. He retains the advantages of Fourier theory and overcomes its shortcomings.

Land, application in print output

WAVELET image compression format is not yet mature, but it is too early as a print output. However, the potential for follow-up development is endless. Especially for online publishing, its use value is even higher. The emergence of WAVELET is just like the emergence of JPEG at the time. There is a whirlwind in the field of imaging, but WAVELET has the advantage that JPEG does not exist. JPEG It is distortion compression, and the degree of recovery after decoding is limited, and it can be applied on the Internet. It is because the resolution of the computer does not need to be too high to recognize the graphics. The resolution needed for printing needs a certain degree. Although WAVELET is also distortion compression, but after decoding can restore the data to almost complete restore, such compression has the value of existence.

One thing that must be raised is that it is not only the restoration of data that can be used in printing, but also the RIP that interprets its files before it can be used for digital printing. Waiting until WAVELET's application matures and then develops its applicable RIP will be a matter of time.

There have been browsers on the Internet that can read WAVELET files. However, they are still beta versions. However, they will be used on the Internet in the future and should be the trend in the future. There should be no small impact on online publishing.

The advantage of image compression lies in the rapid data transmission, the reduction of network usage costs, and the increase of corporate profits. As the time for copying is reduced, it also increases the likelihood of prints being printed locally, reducing shipping costs, reducing expenses, and improving timeliness. Create new business opportunities.

柒, Conclusion

WAVELET's theory is not quite complete, but according to the existing research report, there is still a certain distance from the stage of universal application. However, wavelet analysis has its application value in signal processing, image processing, quantum physics and nonlinear science. There is a formal paper in China to study this compression mode. However, there are many terms that have not been officially translated and each has its own translation. Therefore, it is hard to study. But I believe there will be an official name soon. This also shows that the speed of domestic research is far behind the foreign countries. Many related websites have been set up in foreign countries, and there are only a few related papers in China. As a result, there are still other ways to make this kind of compression model popular. Formal use in the printing industry is a considerable time. However, there is still a considerable opportunity for online publishing, and domestic development is still possible in this area. From the perspective of using its achievements, the printing industry may not need to understand its profound mathematical theory. However, in terms of application, in order to be easy to use and to estimate its development trend, the basic concepts of image compression cannot be eliminated. This article simply introduces one of the image compression modes. The purpose is to make the followers have a basis for reference. Perhaps this mode will become mainstream in the near future, and it will not be at a loss until then.

references:

1. Geoff Davis, 1997, Wavelet Image Compression Construction Kit,.

2. Zhang Weigu. Small Space Studio, First Edition 1994, Image Archive. WINDOWS Implementation (top), Feng Information Co., Ltd.

3. Zhang Weigu. Small Space Studio, First Edition 1994, Image Archive. WINDOWS Actual (below), Feng Information Co., Ltd.

4. Shi Weiming Research Office, 1994, PC image processing technology (2) Image file compression sequel, Flag Standard Publishing Co., Ltd.

5. Lu Yongcheng, Minh, 87. Using the wavelet transform and its application in image and video coding, the private master degree thesis of the Department of Electrical Engineering of Chung Yuan Christian University.

6. Jiang Junming, Min. 86. Introduction to wavelet analysis. Master's thesis of the private Tamkang University's Department of Physics.

7. Zeng Yuyu, Chen Zhuozhou, Minnan 83, the latest digital signal processing technology (voice, image processing practice), Quanxin Information Books.

appendix:

Embedded Zerotree Wavelet Transform, Hierarchical Embedded Zerotree Wavelet Transform, Hierarchical Image Transmission and Progressive Image Transmission

At present, the most commonly used static image compression mode on the Internet is JPEG format or GIF format. However, the amount of data encoded by these formats is not changed, and the receiver must completely accept all the data before displaying the complete image transmitted by the encoder. This phenomenon most often occurs when using the Internet to link WWW sites. We often receive text first. The graphics on the web pages are displayed slowly and in small portions. Sometimes the Internet is seriously congested. After the graphics are displayed only a little bit, it takes a very long time before they are displayed again. It may even be disconnected. This makes the user completely unaware of what graphics are being received. This inevitably wastes network resources. . This shortcoming can be improved by using Embedded Zerotree Wavelet Transform (EZW).

The main function of the hierarchical video transmission system is to allow different size display devices or decoders to obtain signals from the same encoder that meet their requirements. Therefore, it is not necessary to use different encoders for different decoder designs and use them. Increase the scope of its application, and reduce the complexity of the erection system, it can also save more equipment costs. When Shapiro's embedded zerotree wavelet transform (EZW) technology was used to design a hierarchical video delivery system, the encoding effect was not very good. The main reason is that the encoders designed using (EZW) technology are encoded according to the full resolution of the image, which makes it impossible for the decoders with different resolutions and bit rate requirements to be shared by the encoder at the same time. The stream of bits. Although Simultaneous (Simulcast) technology can be used to overcome this, but the technology for the same image with different resolution independent encoding, will make the common lowpass subband (Lowpass Subband) is repeatedly encoded and transmitted, resulting in A very high Redundancy.

Based on the above situation, some people have modified the Embedded Zerotree Wavelet Transform (EZW) technology to complete a new hierarchical video delivery system. The technology is a hierarchical embedded Zerotree Wavelet (LEZW) technology. This technology allows us to design a hierarchical video transmission system that can pre-specify the number of layers and each layer of images before transmission. The resolution and code rate.

The LEZW technique is to extend the continuous approximation quantization (SAQ) in the EZW technology, and the traditional practice of the EZW is to apply the SAQ to all the wavelet conversion coefficients. However, in LEZW technology, SAQ is only used for the coding of one layer (Layer) at a time from the base layer (base layer) until the highest-order layer. When the coding rate of the coded layer is used, it means that the coded layer can be coded in the next layer. In order to improve the efficiency of LEZW, SAQ results in lower layers are applied to higher layer SAQ processes. Based on this coded program, LEZW algorithm can be reconstructed under the limit of the average code rate of each layer. Different resolution images. Therefore, LEZW is very suitable for designing hierarchical video transmission systems.

LEZW technology can also be applied to progressive transmission. For a progressive video transmission system, controlling the resolution will improve the visual quality of the reconstructed image. Commonly used progressive transmission methods use vector quantizers or nulltree data structure encoding algorithms. However, the vector quantizer requires large memory and error sensitivity in transmission and transmission. However, using the progressive video transmission system designed by EZW technology can improve these disadvantages and therefore enjoy better performance. However, it also has the disadvantage that when it is applied to progressive transmission, it is coded and transmitted according to the full resolution. Therefore, when the rate is limited, displaying images with full resolution will make the image blurred. Therefore, when the image at the low bit rate is displayed with a low resolution, the sharpness of the image can be improved.

Therefore, to extend the LEZW technology to progressive transmission, the resolution of each stage (Stage) and the cumulative accumulated rate of transmission of each stage can be set before encoding, and then the code can be encoded and transmitted. The system uses low resolution to display images at low code rates, and displays images at high resolutions at high code rates, improving the visual quality of progressive delivery. In the process of code transmission, the system allows the transmitted bit stream to be interrupted and stopped at any point, and the receiving end can reconstruct the image from the received data, at the resolution of the data interruption.

The design method for progressive video transmission and hierarchical video transmission is similar, except that there are considerable differences in the transmission methods. In a hierarchical video delivery system, all layers of data are sent together in parallel. Progressive video transmission is transmitted in a Stage-by-Stage way. Therefore, the progressive transmission designed using LEZW technology can be viewed as a single-layer system, and its resolution and transmission can be controlled. This waste of Internet resources can be solved to some extent.