About JPEG2000 background

In the past decade, people have witnessed dramatic changes in the storage and transmission of digital images. Digital cameras and inexpensive scanners have rapidly expanded the use of digital images for customers and businesses. In addition, the low cost of preparing PCs (personal computers) for accessing the Internet and multimedia enables digital images to become part of everyday life for many people.

These factors, together with the wide availability of color printers, require image resolution suitable for print reproduction. The image file size necessary to meet these requirements presents important challenges to image storage and transmission. Therefore, digital image compression is actually critical to every aspect of imaging technology, depending on the level of customer, business or scientific applications.

The current JPEG compression (because it is different from JPEG2000) is based on a decision that was made by the JPEG committee in 1988 to use DCT (discontinuous sine transform) combined with Huffman coding as its ISO 10918- The coding method in 1 is now the first part of the multipart standard for still image compression.

The DCT is a lossy compression method that the user uses to define the desired copy quality and therefore obtain a compression ratio that depends on the actual content of the encoded image. This is obtained by applying transformations to the image bitmap. The quantization of pixel depth is a parameter in the transformation. Low-quality images have a limited level of quantization and the result is a very good compression ratio using Huffman coding.

The main benefit of this quality-related coding is that when multiple images are provided at the same time, they are characterized by the same level of quality, both on the screen and at the time of printing.

In 1988, when the DCT algorithm was selected, it was very complicated to implement because of the limited capabilities of PCs and workstations for imaging. JPEG has a difficult beginning, which creates more random interoperability between different implementations of the standard.

In 1992, the chief engineer of a U.S. company at the time suggested in the JPEG committee that a simplified implementation of the standard was proposed with limited color space, algorithms, and other parameters, especially the strict file format specification (JFIF). This was immediately welcomed and became the only JPEG implementation in the market.

JFIF refers to today's JPEG file format. However, this limited tool is somewhat outdated and it is difficult to meet the requirements of high resolution, true color, and additional functionality.

Since the initial introduction of JPEG and JFIF, many additional file formats and compression techniques based on the use of DCT have been introduced. These include SPIFF (Still Image Interchange File Format), JTIP's multi-resolution format (JPEG tiled image pyramid), and newer tools called Flashpix's same concept.

JPEG2000

By 1995, the general computing power of computers had increased, and it was possible to have software decoders that display images as fast as or faster than JPEG-effective hardware implementations.

In 1996, the JPEG committee decided to study the possibility of more powerful compression of image data, in particular the possibility of fractals and wavelet technology providing new solutions. Because of these advances in research, it seems that by the time the new standard was ready in 2000, the name JPEG2000 was chosen.

The goal is to create an advanced, standardized image coding system to serve the next century of applications. The purpose is:

1. To determine the current standard does not produce the best quality or performance range;

2. Provide possibilities for markets that do not currently use compression;

3. Provide an open system method for imaging applications.

Different intent applications have different requirements. E.g:

â–³ Internet applications (global network imaging) must be improved in quality and resolution and have fast decoding capabilities.

â–³ Mobile applications produce the requirements for error resilience, low power, and sequential decoding, while e-commerce requires image confidentiality and digital watermarking.

Digital photography requires low complexity and compression effectiveness.

â–³ Hardcopy color fax, printing and scanning require removal or tiling.

â–³ Typing digital library/archive applications is a need for metadata and content management.

For remote sensing, multi-element, fast coding and valuable coding areas are very important.

Medical applications are not just a large part and require a variety of lossy and lossless codes.

Approximately 20 recommendations were made for technical requirements, most of them based on wavelet technology. And tested the enhancement of DCT encoding, too many requirements for JPEG2000 and DCT encoding seems impossible to satisfy them.

Wavelet technology was finally chosen as the new algorithm for JPEG2000 with the ability to allow scalability in both resolution and quality. Although the complexity of the wavelet transform depends on the size of the screening program and the use of floating points compared to the integer filter program, the wavelet transform is generally more complex than the current block-based DCT transform calculation. As a full-picture conversion, the wavelet transform also requires more memory than the DCT.

However, a program-based implementation can reduce memory requirements. It should be noted that although the wavelet transform is more computationally complex than the DCT transform, its complexity is standardized by the expected computational power of the computer when the standard is expected to complete, and the wavelet transform is no more complex than the DCT transform when publishing the original JPEG. Therefore, the JPEG committee felt that the present and expected advancements in computer capabilities have reduced the impact of these problems and that the complexity of the encoding-decoding requirements does not appear to be an obstacle to the widespread implementation of wavelet technology.

The scalability of the resolution means that the image code is a continuous code stream, and various levels of resolution are obtained from the code stream as a function of the extent to which the code stream is reached. (That is, only the code stream is transcoded and given a certain level of resolution, the code stream is transcoded continuously for resolution).

The same image file can be used for different applications. Note that this is very interesting: once compressed, the compressed data can be reordered, so it can be decompressed in lines, sequentially by size or sequentially by resolution.

Quality is not only a resolution problem, but also a problem of the quantization of a pixel depth value. The operator who codes the image code determines the level of scalability by resolution or quality. Step by step in accordance with the required quality, translation quality scalable file code, including all lossless results if encoded as such.

This can achieve an interesting feature: the ability to use lossless coding of the image and allow downstream processing to determine the degree of decoding required by the application. This ability is quite interesting for medical imaging and publishing applications.

Standards and where to continue to be effective

JPEG2000 refers to all parts of the ISO15444 standard. The current plan requires the following seven parts:

Part 1. JPEG2000 Image Coding System

Part 2. Extensions (Adding more features and sophistication to the core definition of Part 1)

Part 3. Motion JPEG2000

Part 4. Consistency

Section 5. Reference Software (currently includes Java and C implementations)

Section 6. Composite Image File Format (for File Scan and Fax Applications)

Part 7. Minimum Support for Part 1 (technical report).

By the year 2000, Part 1 is expected to become a fully accredited ISO standard. Part 1 defines the core compression techniques and "minimum file format," and Part 1 uses a scalable file format architecture and the smallest features expected to meet 80% of the market's requirements.

Sections 2 through 6 define the compression and file format extensions. The basic file format, called JP2, is an optional part of the standard, but any implementation of the file format must conform to the standard and read all legal JP2 files. The file contains extensions, but these extensions do not prevent Part 1 from reading program execution section 1 functions. This ensures customer interoperability. The JPX file defined in Section 2 requires an extended reader and the goal is a dedicated market.

The JPEG2000 code stream, regardless of the file format used, is mandatory for all applications. The JP2 file format is the default interface for applications that use the JPEG2000 code stream.

It is important to point out that compression and file formats are separate. Although most applications will use one of the standard file formats when exchanging data compressed with JPEG2000, compressed code streams can be used independently of the file format.

In many cases, the file format used to encapsulate compressed image data is a key issue when it comes to interoperability. The current JP2 file format comes from a joint operation with the digital imaging group DIG2000. In this launch, DIG member companies cooperated with the JPEG committee in the document format proposal to be considered by the creation of the JPIE committee. This work specifies the file formats that various types of ordinary users can use. Its goal is to create hardware and software developers and web developers using the existing JPEG2000 compression standard, an online image database.

Various format technologies have been considered, ranging from TIFF-based methods to Apple-based QuickTime methods. All of these methods mix different levels of complexity and ease of using binary data characteristics. The final choice uses a binary container based partly on Apple's QuickTime file format and MPEG-4 file format. Techniques used to generate this file format include ICC, XML, JPEG2000 compression, and image metadata.

In addition to these, the JPEG2000 committee has considered that many groups want to hide the JPEG2000 bitstream in another file format. Therefore, the problem is to provide appropriate levels of image data and metadata in the bitstream and JPEG2000 file formats to satisfy various uses.

The JP2 file format defines the method of determining how colors should be interpreted. Part 2 (JPX) will specify additional methods and the method between the two parts based on the distinction of complexity. In Part 1, the JP2 file format, with sRGB and gray sRGB (neutral sRGB from sRGB) are required by all readers. In addition, all JP2 readers must support the interpretation of colors using legal ICC profiles. These profiles must comply with Three-ComponentMatrixbased and MonochromeInputProfiles (three-element matrix-based and monochrome input profiles). This provides low flexibility but high interoperability. Although other color spaces suitable for JPX files will be listed, there is no need for support in the JP2 reader. In addition, Part 2 specifies registration procedures suitable for the color space other than those specified.

The JP2 file format defines a file as a series of logical boxes containing image data or metadata (data about image data). This makes it easy to analyze through a box or add a new box to the file structure.

The standard logic boxes for JP2 files are file feature tags, file types and compatibility information, general header information, color palettes, color space specifications, element sorts, and code streams. In addition to the fact that the signature and file type must be first and the header information box must be generated before the code flow, the logical box order is not determined. The vector-determined box is always anywhere and the unknown box is ignored by the standard reader.

Other features to include

Metadata: Metadata is defined as data that carries bits with the image. The purpose is to improve the global file or system that contains the data.