JPEG2000

masonchung

JPEG2000是基於小波變換的圖像壓縮標準，由Joint Photographic Experts Group組織創建和維護。JPEG2000通常被認為是未來取代JPEG（基於離散餘弦變換）的下一代圖像壓縮標準。JPEG2000文件的擴展名通常為.jp2，MIME類型是image/jp2。

% o' f, z/ W; {+ L& E1 ~JPEG2000的壓縮比更高，而且不會產生原先的基於離散餘弦變換的JPEG標準產生的'blocky and blurry'artifacts。JPEG2000同時支持破壞性資料壓縮和非破壞性資料壓縮。另外，JPEG2000也支持更複雜的漸進式顯示和下載。

( ]$ N8 B( V' Z' MJPEG2000是國際標準化組織（ISO）發佈的標準，文檔代碼為ISO/IEC 15444-1:2000。雖然JPEG2000在技術上有一定的優勢，但是到目前為止（2006年），網際網路上採用JPEG2000技術製作的圖像文件數量仍然很少，並且大多數的瀏覽器仍然沒有預設支持JPEG2000圖像文件的顯示。但是，由於JPEG2000在非破壞性壓縮下仍然能有比較好的壓縮率，所以JPEG2000在圖像質量要求比較高的醫學圖像的分析和處理中已經有了一定程度的廣泛應用。


0 j7 T- W6 O' R( L# _; r, {相關標準
, M) [) E! E3 |8 L  a已經發佈的JPEG2000標準包括ISO/IEC 15444-1:2000。另外，和JPEG2000相關的一些額外標準也正在制定和討論中，比如JPEG2000安全圖像傳輸（JPSEC）以及基於連接的JPEG2000圖像瀏覽（JPIP）等。
) S' G  d( }  D9 B6 d7 ^; N
; J% L8 C* `! v+ A: kSeveral additional parts of the JPEG 2000 standard exist, some of them are not yet officially released. Amongst them are ISO/IEC 15444-2:2000, JPEG 2000 extensions, featuring for example trellis quantization, an extended file format and additional color transformations, ISO/IEC 15444-4:2000, the reference testing and ISO/IEC 15444-6:2000, the compound image file format, allowing compression of compound text/image graphics. Extensions for secure image transfer, JPSEC, and connection-based image browsing, called JPIP are currently under discussion by the ISO.
1 Z  W% b* q. B4 q% B" E
/ P! i8 T  h( ]) y/ P# I9 b技術探討
" w1 T2 L  s  t: n) X" uImage:Jpeg 2000.jpg
' d7 x' P; o5 Y! oComparison of JPEG 2000 with the original JPEG format.JPEG 2000的目標不僅僅是性能要超越JPEG，而且增加和增強瞭如可縮放性和可編輯性這樣的特性。
7 E, I! ]& |  W* v! X% e
In fact, JPEG 2000's improvement in compression performance relative to the original JPEG standard is actually rather modest and should not ordinarily be the primary consideration for evaluating the design. Moreover, very low and very high compression rates (including lossless compression) are also supported in JPEG 2000. In fact, the graceful ability of the design to handle a very large range of effective bit rates is one of the strengths of JPEG 2000. (For example, to reduce the number of bits for a picture below a certain amount, the advisable thing to do with the first JPEG standard is to reduce the resolution of the input image before encoding it — something that is ordinarily not necessary for that purpose when using JPEG 2000 because of its wavelet scalability properties.)
0 `* ^5 H& f  [/ U
JPEG 2000, as did the original JPEG 1992 standard, applies a form of transform coding to compress images. However, JPEG 2000 uses a wavelet transform, in contrast to JPEG 1992 which uses an 8x8 block-size discrete cosine transform.

' `2 S4 s% O' n) T) b' ]Initially, images have to be transformed (from the RGB color space) to the well known YCbCr color space or to the RCT space (reversible component transform) leading to three components. In lossless mode a very rough, but reversible, approximation of this color space conversion is used. The chrominance components can be, but not necessarily have to be, down-scaled in resolution; in fact, since the wavelet transformation already separates images into scales, downsampling is more effectively handled by dropping the finest wavelet scale. This step is called multiple component transformation in the JPEG 2000 language since its usage is not restricted to the RGB color model.
8 \0 x' n$ u8 y% Z/ o; i# F7 G
After color transformation, the image is split into so-called tiles, rectangular regions of the image that are transformed and encoded separately. The purpose of tiles is to cope with memory limitations more easily. These tiles are then wavelet transformed to an arbitrary depth.
7 B+ M& J7 h$ V! S
$ b' U+ d/ r; K2 w! Q1 q1 x, ^The result is a collection of sub-bands which represent several approximation scales. A sub-band is a set of coefficients — real numbers which represent aspects of the image associated with a certain frequency range as well as a spatial area of the image. These coefficients are scalar-quantized, giving a set of integer numbers which have to be encoded bit-by-bit.
$ I* D' ?7 ]5 W2 }/ b
' S) H7 I* _% v' XThe quantized sub-bands are split further into precincts, rectangular regions in the wavelet domain. They are typically selected in a way that the coefficients within them across the sub-bands form approximately spatial blocks in the (reconstructed) image domain, though this is not a requirement.
4 y( m* f' H$ n' `
5 ~% H5 Q1 X5 l+ ^6 j- ]Precincts are split further into code-blocks. Code-blocks are located in a single sub-band and have equal sizes — except those located at the edges of the image. The encoder has to encode the bits of all quantized coefficients of a code-block, starting with the most significant bits and progressing to less significant bits by a process called the EBCOT scheme. EBCOT here stands for Embedded Block Coding with Optimal Truncation. In this encoding process, each bit-plane of the codeblock gets encoded in three so called coding passes, first encoding bits (and signs) of insignificant coefficients with significant neighbors (i.e. with 1-bits in higher bit-planes), then refinement bits of significant coefficients and finally coefficients without significant neighbours. The three passes are called Significance Propagation, Magnitude Refinement and Cleanup Pass, respectively.
6 e# ~2 t. c  @3 [' A' ]$ m# Z
4 y" X- l3 l& _1 D# p- ^The bits selected by these coding passes then get encoded by a context-driven binary arithmetic coder, namely the binary MQ-coder. The context of a coefficient is formed by the state of its nine neighbours in the codeblock.
: u% |/ ?& c% q& m8 j; ^1 ^  M
9 S1 j4 D7 y  P8 u( S# rThe result is a bit-stream that is split into packets where a packet groups selected passes of all codeblocks from a precinct into one indivisible unit. Packets are the key to quality scalability (i.e. packets containing less significant bits can be discarded to achieve lower bit-rates and higher distortion).

Packets from all sub-bands are then collected in so-called layers. The way the packets are built up from the code-block coding passes, and thus which packets a layer shall contain is not defined by the JPEG 2000 standard, but in general a codec will try to build layers in such a way that the image quality will increase monotonically with each layer, and the image distortion will shrink from layer to layer. Thus, layers define the progression by image quality within the codestream.

The problem is now to find the optimal packet length for all code-blocks which minimizes the overall distortion in a way that the generated target bitrate equals the demanded bitrate. While the standard does not define a procedure as to how to perform this so-called rate-distortion optimization, the general outline is given in one of its many appendices: For each bit encoded by the EBCOT coder, the improvement in image quality, defined as mean square error, gets measured; this can be implemented by an easy table-lookup algorithm. Furthermore, the length of the resulting codestream gets measured. This forms for each codeblock a graph in the rate-distortion plane, giving image quality over bitstream length. The optimal selection for the truncation points, thus for the packet-build-up points is then given by defining critical slopes of these curves, and picking all those coding passes whose curve in the rate-distortion graph is steeper than the given critical slope. This method can be seen as a special application of the method of Lagrange multiplier which is used for optimization problems under constraints. The Lagrange multiplier, typically denoted by λ, turns out to be the critical slope, the constraint is the demanded target bitrate, and the value to optimize is the overall distortion.

5 m3 w: s/ i( A4 ~- H6 k6 l  I5 sPackets can be reordered almost arbitrarily in the JPEG 2000 bit-stream; this gives the encoder as well as image servers a high degree of freedom.

6 o: z% h* J0 t6 MAlready encoded images can be sent over networks with arbitrary bit-rates by using a layer-progressive encoding order. On the other hand, color components can be moved back in the bit-stream; lower resolutions (corresponding to low-frequency sub-bands) could be sent first for image previewing. Finally, spatial browsing of large images is possible through appropriate tile- and/or partition selection. All these operations do not require any re-encoding but only byte-wise copy operations.
( E* e% u8 h) O& O1 f" l
# f2 ~8 s5 K# D0 T0 x* h9 }JPEG 2000 gains up to about 20% compression performance for medium compression rates in comparison to the first JPEG standard. For lower or higher compression rates, the improvement can be somewhat greater (especially if altering the input resolution to the codec is not considered as a technique for effective use of the older JPEG standard). Good applications for JPEG 2000 are large images, images with low-contrast edges — e.g. medical images.

$ L" A* M8 N+ vIt has, however, notably higher computational and memory demands.
) i3 \' c( L" [9 W* b2 `
) _7 y8 U5 x& ^' FLossless compression is achieved through the use of a special integer wavelet filter (biorthogonal 3/5 instead of Daubechies biorthogonal 7/9) and a quantization step size of 1. Clearly, in lossless mode all bitplanes have to be encoded by the EBCOT, and no bitplanes can be dropped.

. U& z3 |: G9 ]7 b[ 本帖最後由 masonchung 於 2007-1-20 02:42 PM 編輯 ]

duffrose

不知道這個東西合不合適放在硬體裡面？目前的狀態是支援的軟體不多的樣子，在網路上瀏覽的照片還都不是JPEG2000,是軟體支援程度有差還是授權的問題呢？

masonchung

有一些學校已經做出 jpeg2000的晶片嘞
% f7 i- c  W* `+ u在網路上瀏覽的照片還都不是JPEG2000 主要是壓縮軟體支援不普及