Research On Image Coding Based On Wavelet Transform

With the continuous development of computer and network technology, theextensive used of multimedia technology, the related technologies of wirelesssensor and the Internet of Things to be improved, the image has become anintegral part in the exchange of information and transmission process. Theshorten replacement cycle of electronic products and hardware, and thefacilitate of the frequent technical innovations are changing the life, at thesame time, they lead to rise sharply in the amount of information contained inthe image information, and to directly limit the image and its related productsreleased, transmitted and stored in the network, also caused the "InformationExplosion". By compression coding the image information can be a goodsolution to this problem. Study and explore efficient image codingcompression algorithms or coding framework naturally become a hot topic.For the multi-resolution characteristics and local characteristics wavelettransform technology has been widely studied and applied.Firstly, briefly the theoretical knowledge of the Wavelet Transform,analysis the strengths and weaknesses of the Wavelet Transform. About theshortcomings of WT, introduce a new wavelet transform–Dual-Tree DiscreteWavelet Transform (DDWT). Subsequently, deathly analysis the correlationproperties of the wavelet coefficients of DDWT, emphatically discuss how tosparse the wavelet coefficients. Finally introduce the algorithm of StructuralSimilarity Index Noise-Shaping to reduce and sparse representation of thecoefficients.Second, elaborate the significant organizational structure in the field of theimage coding algorithm based on wavelet transform: Zero-Tree structure;briefly descript the Embedded Zero-Tree Wavelet algorithm (EZW) and theX-Tree coding algorithm. About the EZW coding algorithm insufficient,according to the characteristics of Dual-Tree Discrete Wavelet Transformcoefficients, propose a kind of Similar-Wavelet-Packet structure Enhanced X-Tree coding algorithm. The first stage is that wavelet coefficients aredivided and pocketed; the next stage is that reorganize the pocketed waveletcoefficients into the structure of the similar wavelet packet structure; the laststage is that take advantage of the Enhanced X-tree structure to code theDDWT wavelet coefficients, which is better efficiency than EZW coding.Finally, study the SPIHT coding algorithm, dissect the merits anddrawbacks of the algorithm, and propose to combine the partitioned sub-blockwith Multi-Channel Parallel to encode, and then utilize the Space OrientationTree in all directions after the DDWT wavelet transform to tissue waveletcoefficients, and to achieve the improvement of the algorithm.