The Research And Improvement On SPECK Image Coding Algorithm

Withthewideapplicationofnetworks andtheneedofwirelessvideotransmission,the processing capacity and transmission capacity of the image data increase sharply.However, the existing hardware facilities for the image processing and transmission arehard to keep pace with its development, which brings a great burden on communicationbandwidth and storage medium. It becomes an inevitable choice for the image data toadapt codingcompressiontechnologyforprocessingtoreducetransmissioninformation.To some degree, the image compression coding has become an urgent subject.Embedded coding is getting more and more popular in image increasing transmission,image brown and image transmission in networks, due to its special properties such astheavailabilitytoincreasingtransmission,theabilitytoceaseencodingat anypoint andtherigidsatisfactiontotargetcoderatioortargetdistortion.Among so many embedded image coding algorithms, the EZW, SPIHT andSPECK are the best representatives. The EZW algorithm is regarded as a great leap, ituse a"zerotree"data structure to realize the effective organization of the waveletcoefficients. The SPIHT algorithm causes a better performance by improving theEZW's data structure. The SPECK algorithm makes full use of the characteristics ofwavelet coefficients involving the energy clustering and the energy decreasing alongwith the increase of scalability. Furthermore, by combining the quadtree partition withthe bit-plane encoding, it gains a good compressing performance and more attention.However,therestill exists someimprovement tobedone as fortheachievingspeedandcomplexity, in spite of its coding independently, fast coding speed and so on. Animproved SPECK algorithm in the sequence of process and the importance test and etcis presented in this dissertation. Besides the high coding efficiency, the improvedalgorithm preserves the properties of the SPECK. The experimental results show thatthe codingefficiencycan be further improved and the computational complexitycan beloweredbytheimprovedSPECKalgorithm.At present, the image compression technology based on wavelet transform is stillin constant development. A lot of questions are looked forward to further research andperfection,whichwillbeconductedinourfuturework.