This thesis gives an overview of entropy coding in jpeg 2000 , especially the code - block based coding and context based adaptive arithmetic coding . to optimize the implementation of mq encoder , several methods are adopted to optimize the algorithm which is emulated in matlab 因此本文對(duì)jpeg2000標(biāo)準(zhǔn)中位平面編碼和基于上下文的自適應(yīng)二進(jìn)制算術(shù)編碼技術(shù)作了詳細(xì)的分析和研究之后,對(duì)自適應(yīng)算術(shù)編碼算法優(yōu)化使其更適合于硬件實(shí)現(xiàn)。
Here we introduce the unique characteristic of h . 264 in detail , which includes intra frame residual coding , integer cosine transformation , binary adaptive arithmetic coding , concept network adaptation layer , an ember of different block sized used for motion prediction and 1 / 8 pixel motion estimation etc 摘要介紹了h . 264獨(dú)有的特點(diǎn),包括幀內(nèi)預(yù)測(cè)、整數(shù)余弦變換、二進(jìn)制自適應(yīng)算術(shù)編碼、概念性網(wǎng)絡(luò)適配、多種塊尺寸運(yùn)動(dòng)搜索、 1 / 8像素運(yùn)動(dòng)估計(jì)精度等。
In this thesis , we review the joint source and channel coding firstly , introducing fundamental theories , design methodologies , and practical applications . secondly , an approach is presented for improving arithmetic coding by embedding hamming distance . the approach prevents error propagation of the arithmetic coding 作者在對(duì)算術(shù)編碼做了深入研究的基礎(chǔ)上,運(yùn)用漢明距離對(duì)其加以改進(jìn),使算術(shù)編碼具有了較好的糾錯(cuò)能力,克服了算術(shù)編碼易受傳輸錯(cuò)誤的影響,使其適用于信道干擾較大的場(chǎng)合。
Neural networks are used more frequently in lossy data coding than in general lossless data coding , because standard neural networks must be trained off - line and they are too slow to be practical . in this thesis , statistical language model based on maximum entropy and neural networks are discussed particularly . then , an arithmetic coding algorithm based on maximum entropy and neural networks are proposed in this thesis 傳統(tǒng)的人工神經(jīng)網(wǎng)絡(luò)數(shù)據(jù)編碼算法需要離線(xiàn)訓(xùn)練且編碼速度慢,因此通常多用于專(zhuān)用有損編碼領(lǐng)域如聲音、圖像編碼等,在無(wú)損數(shù)據(jù)編碼領(lǐng)域應(yīng)用較少,針對(duì)這種現(xiàn)狀,本文詳細(xì)地研究了最大熵統(tǒng)計(jì)語(yǔ)言模型和神經(jīng)網(wǎng)絡(luò)算法各自的特點(diǎn),在此基礎(chǔ)上提出了一種基于神經(jīng)網(wǎng)絡(luò)和最大熵原理的算術(shù)編碼方法,這是一種自適應(yīng)的可在線(xiàn)學(xué)習(xí)的算法,并具有精簡(jiǎn)的網(wǎng)絡(luò)結(jié)構(gòu)。
Secondly , programmed the image processing arithmetic code which include the bottom arithmetic for the general condition comprises threshold division , region combination and informate and the middle level arithmetic for the given task comprises detecting the line dation creirection according to the hough transform in order to fix on the hole ’ s azimuth angle , detecting the aiguille tip position according to the image movement according to the environment and the image format 然后,根據(jù)目標(biāo)環(huán)境要求和攝像機(jī)采集圖像格式,開(kāi)發(fā)了圖像處理算法程序。圖像處理算法包括底層算法和中層算法兩部分,底層算法針對(duì)通用情況,包括閾值分割、區(qū)域合并和信息生成。中層算法針對(duì)具體任務(wù)設(shè)計(jì),包括利用hough變換檢測(cè)棱線(xiàn)的方向,從而確定圓孔的方位角和利用基于圖像運(yùn)動(dòng)檢測(cè)鉆頭尖端位置。
Text is a kind of very common resource in digital library , and lossless techniques play an important role in compressing text . starting from the shannon ' s entropy theory , we analyze the lossless compression algorithms , and implement arithmetic coding algorithm in c . in the experiments , we compare four different lossless compression algorithms by their performances such as compression rate , compression rate tendency with the length of data , stability , and complexity , using 35 groups data series with 4 different length 本文從信息論中shannon熵定理出發(fā),對(duì)無(wú)損壓縮技術(shù)進(jìn)行系統(tǒng)地分析,用c語(yǔ)言實(shí)現(xiàn)了其中的算術(shù)編碼算法,并用它對(duì)35組、四種不同長(zhǎng)度數(shù)據(jù)序列進(jìn)行了壓縮,給出了實(shí)驗(yàn)結(jié)果,然后從壓縮比、壓縮比隨字符串長(zhǎng)度的變換趨勢(shì)、算法穩(wěn)定性和算法復(fù)雜性等四個(gè)方面對(duì)其與其它三種壓縮算法lzw 、 lz77 、 rle進(jìn)行了分析與比較。
Arithmetic coding is a form of entropy encoding used in lossless data compression. Normally, a string of characters such as the words "hello there" is represented using a fixed number of bits per character, as in the ASCII code.