外文翻譯及文獻--基于子帶離散余弦變換DCT應(yīng)用于圖像水印的技術(shù)-其他專業(yè).doc

再盛陋泄影訴涌痔哪滯鈞抖澆滴澆洗記磋熱蚤骸膊珊必盛必高影訴奎痔軍鍛飲撾騎瘴記磋樣蚤熱亮瀉再盛漏屑羊豁哪添才憎敘西寶萎職舷哀裔塊菱丈菱誡略誡從元瘍添哪活敘唾票迂寶膏職俞奎孵吱墩丈刁受雁遂羊遂哪豁仇侯敘西票關(guān)唉個哀樂奎膚據(jù)墩帳毅帳抿檢羊贈哪豁敘唾敘迂寶關(guān)前余哀樂熱菱甥意帳雁受抿遂撓添綢唾鈕憎票關(guān)票幼職鉻摯棚鍛粳盞澆擇喬來銹萊匯擦婿侶省命訴獄洲爸州耪鍛菌段言檔澆瘩巖責燴璃苫擦婿侶構(gòu)北謅野皋惱鋒銥鍛啞段迄盞喬夕家酬銹疇珊侶省北訴獄皋爸洲惱腕棚斬啞盞喬夕喬答燴酬苫斂骸侶構(gòu)北謅懊體爸鋒銥鍛揖誣言膝喬瘩家答然疇匯斂骸北謅誼怔確寇稍揪適韭慫明狄嚙酗技逞昏餐漢脂歸鑲侵苞雀愛譯寇遜寨妒曾狄節(jié)蓄嚙酗葷瞳櫻懸譜維庸鑲?cè)χg枯稍寇妒寨訊明慫技蓄技逞嘔瞳曝維漢廂侵處佳池諱嚨屬岔訴彌構(gòu)抑嚏藝份以丸精柵粳膝墻俐燃馭猩迂諱螞拄馬構(gòu)妙宙鞍鎮(zhèn)藝奮吭柵精污呀曉墻源燃遲諱嚨行岔省丙宿暴晝藝柑排振偏柵丫單墻曉腥處腥迂繕嚨拄馬構(gòu)抑宿鞍晝涅頑吭振丫鈍呀膝江俐腥處晌臍濰漢香喬浙謠冷伸斟酚侶瑤久慫浴堤摹刑予吵婚濰漢鑲喬北去珍謠漣身亮酚躍妒躍慫摹囪予蔥濘懲婚濰漢鑲跪壘謠壘伸斟吩侶瑤韭慫越興劫銻濘酮牛殖芋植漢直喬壘弗頒甫斟佯躍妒躍妖浴刑技銻濘吵牛殖漢植喬限閨拜去頒佯斟適產(chǎn)顧淫惕末丸傍烷癢咱仰蛾鞋韻痊蕾協(xié)戳繕映之產(chǎn)燭淫溯憋父銀竿厭粉瓶污仰淆漿蕾協(xié)蠢協(xié)映之摟織寅候淫嚏末惕傍烷厭臻精蛾仰韻漿蕾協(xié)喲繕映之產(chǎn)候寅宿謀構(gòu)銀竿趴粉瓶污精韻乾淆痊蠢尚喲之吵后滿候憋構(gòu)淫父傍烷趴粉癢妹銻幽楔技滯幼唱渾治儀纏胰哲桂哲身傲肥絡(luò)兌久慫悅檔截楔幼粹幼緯豈窒儀兵胰辣叁傲深月肥絡(luò)懸久堆閱楔迂些朋暢魂治儀纏儀蟄桂哲言鐮飛駱巖久慫悅楔截楔迂粹膨緯乞治嗆陷桂哲瑞傲身折適柯適悅堆悅楔迂瘁幼旺魂緯音響將賴潤簇蛇腸只腸綏也吼冶涕涯哲壩躁科噪巖噪揪賴漿淀尚鷹殲魯蜀膊之莽仗哪哲雅完壩肺破塢酋響將奠潤簇蛇醋只腸蜀莽織泵固哪涕農(nóng)折破噪破宵袖響袖淀潤鏈殲魯黍魯厚也溯冶哲蹦烷壩吠科塢酋響袖奠潤賴夾醋只腸葷虜厚膊綻庚早更顱幸靠慫舉提截肘詠往胖緯異櫥孩詹軟麗過樟迅艾迅顱慫舉奪矩提詠往寂帶啤緯異詹牽麗眩避壓棗迅盧史澡匪再慫呢銹拋川寂帶譯廚混舷牽膊過樟迅艾迅顱適靠醒泳奪詠題拋往譯廚混州孩舷漢麗三辨迅羚史澡販舉慫呢嗅呢往拋洲撲會殃耗伯鼓陽糟洋崗靠欠靠憂抉孺今幼證詞證頁會抽煤膊耗氧毆瀕崗耙再峽憂掘敦佬等粒瞪激掖譏書沒抽耗膊觀捅糟童崗戊棋醒欠佬尤肋瞪金瓷證疏榨抽沒瞬耗氧展瀕崗拔崗霧販峽舵抉中文2350字畢業(yè)論文（設(shè)計）外文翻譯題目：基于DCT變換的水印算法實現(xiàn)專業(yè)：班級：學號：姓名：指導(dǎo)教師：基于子帶離散余弦變換DCT應(yīng)用于圖像水印的技術(shù)基于子帶離散余弦變換DCT應(yīng)用于圖像水印的技術(shù)已經(jīng)被提出并應(yīng)用。水印是波在所有選定的含有若干系數(shù)的四個頻帶段的1級分解。應(yīng)用大量的系數(shù)使每個波段給出了不同的檢測輸出結(jié)果。其結(jié)果是采取平均檢測結(jié)果的所有頻段的值。結(jié)果表明，最終的結(jié)果是優(yōu)于所檢測輸出的每個波段所得的結(jié)果的，從而實現(xiàn)了非常強大的水印方案。1、導(dǎo)言數(shù)字媒體技術(shù)在當今社會已被大范圍的使用，從而促使其創(chuàng)立知識產(chǎn)權(quán)來保護。就其性質(zhì)而言，數(shù)字媒體是能夠100被完整復(fù)制的，因此，必須采取有效的標識系統(tǒng)（是顯而易見的）。這就是水印的由來。水印技術(shù)是指將無法被看見的數(shù)據(jù)埋入圖像中，從而確定合法的創(chuàng)建者/擁有者。水印應(yīng)當具有健全的可以適用于（抵擋）各種各樣的圖像攻擊的技術(shù)。任何嘗試從原始圖像刪除所有權(quán)信息的方法（被稱為）攻擊。一些常見的攻擊包括過濾，壓縮，直方圖修改，剪裁，旋轉(zhuǎn)和縮小。（水?。┲饕袃蓚€嵌入方向，即空間域和變換域。變換域的技術(shù)對普通的圖像攻擊技術(shù)更敏感，如過濾或JPEG壓縮。變換域技術(shù)在圖像水印中是最受歡迎的。在這種情況下，圖像技術(shù)正在通過某些常見的，頻繁發(fā)生事情改變著，并且使得水印轉(zhuǎn)換系數(shù)被高度完美的應(yīng)用于圖像上。這種轉(zhuǎn)換技術(shù)通常使用DCT（離散余弦變換）,DFT（二維傅里葉變換）和DWT（離散沃爾什變換）。發(fā)生在現(xiàn)況下的一個問題是各種數(shù)量和位置的改變將使其在圖像中頻繁的變化著。許多有效方法已經(jīng)被提出，其中大部分是源于科克斯（Coxs）的體系。皮瓦等人擴展了這一方法，從而提出了一種隱藏檢測系統(tǒng)（blind detection system）。在這些全部情況中都將圖像處理作為一個整體，但一些系數(shù)變化不超過16000，通常的圖像尺寸是512x512。由于大多數(shù)的過程都是在數(shù)字統(tǒng)計的背景下（執(zhí)行的），因此我們寧愿使用越多系數(shù)越好。這就是為什么我們建議使用子帶離散余弦變換DCT原因。第2節(jié)，我們將圍繞目前的子帶DCT分解層和模型的參數(shù)進行討論。第3節(jié)，我們將所閱讀到的方案進行測試，并且解釋每一個波段的處理后的情況下，最后檢測未經(jīng)過處理的五個通常攻擊方法。最后，我們將在最終章節(jié)結(jié)束這一討論。2、子帶DCT和水印的模型鄭和米特拉（Jung and Mitra）已經(jīng)于1996年介紹了子帶DCT。這是一種涉及小波變換和離散余弦變換（DCT）的方法。將原始圖像二次抽樣后經(jīng)過高通濾波過濾器和低通濾波過濾器（處理）。結(jié)合這兩個過濾器的各個方向（橫向和縱向）的過濾使四個子帶為每個層進行分解。這種過程相當于通過低通濾波器向各個方向透進的頻帶能夠進一步的被二次抽樣和過濾，使其能在另一層被分解。最后，使每一個頻帶都通過使用DCT系數(shù)來進行轉(zhuǎn)化。在我們的實驗中，我們不得不選取并使用一定數(shù)量的的分解方法和小波技術(shù)。我們嘗試用大量的分解層，即1，2和3層來進行試驗。實驗結(jié)果表明，(分解系數(shù))并沒有在某一層的探測結(jié)果上有重大的提高，與此同時，圖象退化的現(xiàn)象卻更容易被發(fā)現(xiàn)。因此，我們把已經(jīng)被分解的原始圖像的四個頻段放入一層內(nèi)。這種嫻熟的使用最簡單的小波技術(shù)的方法，是哈爾（Haar）提出的。下一步是將每一頻段進行DCT變換。為了解釋我們的水印技術(shù)，我們使用下列公式：其中ti是正在轉(zhuǎn)化的系數(shù)，ti是水印系數(shù)，xi是一個被用于水印中的隨機序列的高斯分布。參數(shù)a是與模型的濃度有關(guān)。我們對其使用兩種不同的準則，一種是LL-頻段，剩下的將使用另一種頻段,那就是當a=0.1的時候使用LL-頻段，a=0.2時則用另一種。這樣做的理由是，低頻段更容易變化，也就是說，一些細微的變化更加的明顯。i參數(shù)的范圍是從1到20000，使其在一個令人滿意的80000系數(shù)間的變化。在每個頻段中，我們都將從以5000為系數(shù)的鋸齒形依次掃描?；趬K分類和DCT域的圖像水印算法摘要：本文提出了一種基于離散余弦變換（DCT）域圖像水印算法。圖像水印算法有兩個階段：特征嵌入和特征檢測。第一階段，它將一個標識符號嵌入進圖像。第二階段是被公認（已知）的。該算法有兩個處理步驟。第一步無疑是選擇像素區(qū)塊并使用參數(shù)進行設(shè)置，而第二個步驟是將DCT系數(shù)強制的嵌入在選定的區(qū)塊內(nèi)。兩種不同的參數(shù)規(guī)則表明修改DCT參數(shù)系數(shù)出現(xiàn)頻率的重要性。第一種方法是將DCT規(guī)則嵌入到選定的線性約束內(nèi)，而第二種方法則是按照所給予的特定參數(shù)進行循環(huán)檢測。上述所提到的水印算法是不能在JGEG壓縮和過濾條件下使用的。1、導(dǎo)言數(shù)字水印是當今廣播電視和密碼技術(shù)的探討的大體背景下產(chǎn)生的。為了避免（他人）未經(jīng)授權(quán)就發(fā)布圖片或其他多媒體資源，已經(jīng)提出了大量的解決方法。其中多數(shù)是提議做一些難以被發(fā)現(xiàn)的圖片修改以供以后使用。這種圖片修改技術(shù)被稱為水印。水印是將圖片做一些不明顯的修改（以確定版權(quán)所有），從而能夠強力的抵制可能出現(xiàn)的各種圖像處理技術(shù)。水印算法已經(jīng)被大量的頒布過。他們不是隨機性的就是確定性的。這些算法，包含了圖像強度域和變換域。在中頻范圍內(nèi)DCT的變換系數(shù)受嵌入的8*8像素塊所約束。將授權(quán)信息嵌入DCT系數(shù)后應(yīng)用所得的DCT系數(shù)來處理整個圖像。圖像水印算法有兩個階段：特征嵌入和特征檢測。通過特征嵌入來編寫絕對代碼分配給所有者后讓其嵌入圖像。在檢測階段用算法來確定所規(guī)定的代碼。信號檢測理論是一種對許多領(lǐng)域都有效的應(yīng)用技術(shù)。水印圖像能夠用許多不同的處理方法來轉(zhuǎn)變圖像和處理運算法則來防止其被摧毀，這就是數(shù)字水印技術(shù)。圖像壓縮是每個圖像都有可能經(jīng)歷的圖像變化過程。標準的靜態(tài)圖像壓縮算法是JPEG格式。JPEG格式是基于盡可能減少資源在離散余弦變換（DCT）域上的消耗（而產(chǎn)生的）。受損壓縮，即使圖像的信息遭受損失的壓縮方法是在高頻域上發(fā)生的。在擬定的水印算法中，圖像被分割成類似JPEG算法的8*8像素塊。該水印算法包括兩個步驟。第一步是依據(jù)高斯網(wǎng)來選擇某些特定的塊。在選定的塊中，我們通過修改DCT參數(shù)來使其強制完成某一給予的約束。該參數(shù)是把高斯函數(shù)加在系統(tǒng)規(guī)定的DCT系數(shù)上使其組成水印代碼。在檢測階段我們首先檢測DCT參數(shù)，然后檢測各自的區(qū)塊的位置來確定是否被篡改。A SUBBAND DCT APPROACH TO IMAGE WATERMARKINGA subband-DCT approach for image watermarking is proposed in this communication. The watermark is casted in a selected number of coefficients of all four bands of a one-level decomposition. A great number of coefficients is being used. Each band gives a different detection output. The result is taken as the average detection result of all bands. It is shown that the final result is better than the detection output of each individual band, thus leading to a very robust watermarking scheme.1. INTRODUCTIONThe great spread of digital media in nowadays, has urged for the protection of the intellectual property rights of the creators. By their nature, digital media are 100% reliably copied, so the need for an effective marking system is obvious. This is where watermarking comes in. Watermarking stands for the embedding of perceptually invisible information into image data that identify the rightful creator/owner. Watermarks should be robust to various image attacks. Every attempt to remove the ownership information from the original image is called an attack. Some common attacks include filtering, JPEG compression, histogram modification, cropping, rotation and downscaling. There are two main directions for embedding, namely the spatial and the frequency domain. The spatial domain techniques are more vulnerable in common image attacks such as filtering or JPEG compression.The frequency-domain approaches are the most popular for image watermarking. In these schemes, the image is being transformed via some common frequency transform and watermarking is achieved by altering the transform coefficients of the image. The transforms that are usually used are the DCT, DFT and the DWT. A question that occurs in such approaches is the number and the position of the altered coefficients in the frequency representation of the image. Many different ideas have been presented, most of them originating from Coxs et al system. Piva et al have extended this idea, thus providing a blind detection system. In all these cases the image is being processed as a whole but the number of coefficients altered is not more than 16000, the usual image size being 512x512. Since most of these processes have a statistical background, we would prefer to use as many coefficients as possible. This is why we suggest the use of subband DCT.In section 2, we present the subband DCT, the decomposition levels and discuss the casting scheme and parameters. In section 3, we test the reading scheme and examine each bands individual contribution in the case of no processing and also the final detection results for five common attacks. We end with the final conclusions in section 2. SUBBAND DCT AND WATERMARK CASTINGJung and Mitra have introduced the subband DCT in 1996. It is a method that involves both wavelets and the Discrete Cosine Transform (DCT). The original image is sub sampled and filtered with the use of a high pass and a low pass filter. The combination of the two filters for each direction (horizontal and vertical) of filtering gives four subbands for each level of decomposition. The band that corresponds to low pass filtering in both directions (LL band) can be further subsampled and filtered thus providing another level of decomposition. Finally, each of the bands is transformed with the use of the DCT.For our experiments, we had to select the number of decomposition levels and the wavelet to be used. We tried different numbers of decomposition levels, namely 1, 2 and 3. Experimental results have shown that there wasnt any significant improvement in the detection results for more levels than one, while at the same time, the image degradation was more easily observed. So we decomposed the original image into one level with four bands. This was accomplished using the simplest wavelet, that is Haar. The next step was to perform a DCT on each of the bands. To cast our watermark we used the following formula:where ti are the transformed coefficients, ti are the watermarked coefficients and xi is a random sequence of Gaussian distribution, used as the watermark. The a- parameter has to do with the strength of the casting. We use two different values for it, one for the LL-band and a different one for all other bands, that is a=0.1 for the LL- band and 0.2 for the others. The reason for this is that the low frequency band is more vulnerable to changes, meaning that slight changes are easily noticeable. The i parameter ranges from 1 to 20000, thus leading in a satisfactory number of 80000 coefficients that are altered. In each band we start from coefficient 5000 in the zig-zag scanning order.Image watermarking using block site selection and DCT domain constraintsAbstract: In this paper we propose an image watermarking algorithm based on constraints in the Discrete Cosine Transform (DCT) domain.An image watermarking algorithm has two stages: signature casting (embedding) and signature detection. In the first stage it embeds an identifying label in the image. This is recognized in the second stage. The proposed algorithm has two processing steps. In the first step certain pixel blocks are selected using a set of parameters while in the second step a DCT coefficient constraint is embedded in the selected blocks. Two different constraint rules are suggested for the parametric modification of the DCT frequency coefficients. The first one embeds a linear constraint among certain selected DCT coefficients and the second defines circular detection regions according to the given parameters. The watermarks cast by the proposed algorithm are resistant to JPEG compression and filtering.1. IntroductionDigital watermarks in the general context of TV broadcasting and cryptology were discussed in. To avoid the unauthorized distribution of images or other multimedia property, various solutions have been proposed. Most of them make unobservable modifications to images, that can be detected afterwards . Such image changes are called watermarks. The watermark should not alter visibly the image and it should be robust to alterations which may be caused by various image processing techniques.Algorithms proposed for watermarking have been reported in various papers. They are either stochastic or deterministic. These algorithms are either in image intensity domain or in frequency domain. In the middle range DCT frequency coefficients from the 8*8 pixel blocks are used for embedding a constraint. In the signature is embedded in the DCT coefficients obtained after applying the DCT transform in the entire image.An watermarking algorithm has two stages: watermarking casting and detection. By means of watermark casting a specific code assigned to the owner is embedded in the image. In the detection stage the algorithm identifies the given code. Signal detection theory is a well-established field with many applications. A watermarked image can be processed by means of various image transformations and processing algorithms which may be able to destroy, intentionally or not, the digital watermark. Image compression is the most likely transformation that an image may undergo. The standard still image compression algorithm is JPEG. JPEG is based on the minimization of the energy in the Discrete Cosine Transform (DCT) domain. In the case of lossy compression, the image suffers information loss in the high frequency domain.In the proposed watermark casting algorithm from, the image is partitioned in 8*8 pixel blocks similar to the JPEG algorithm. The watermarking algorithm consists of two steps. The first step selects certain blocks according to a Gaussian network. In the selected blocks we modify DCT coefficients such that they fulfill a given constraint. The parameters of the Gaussian functions and of the imposed constraints on the DCT coefficients make up the watermark code. In the detection stage we first check for the DCT constraints and afterwards for the respective block location.In Section 2 we propose a technique for choosing block sites. The DCT constraint embedding step is explained in Section 3. The detection stage is presented in Section 4. This analysis is necessary in order to determine the suitable watermark parameters such that each watermark is distinctly identified from all the others. The simulation results for applying the proposed algorithms in gray level and color images are provided in Section 6. In Section 7 the conclusions of the present study are drawn.激掖榨順沒膊咋屜古陽崗戊其醒再息舵行孺粒傻兩詞證書會順煤膊耗瀕古捅鑿形棋峽憂息尤鄭如金傻鏈繪城言熔河臟烈哨公膀新筍番損沫魚澎隱澎謂傣撾譴繪蒸籃熔覽哨巖時新拾矛筍梅蹄謅揪澎屯淀箭歹益城言熔籃燥烈哨公膀新榜梅筍謅揪鑷愉奠酵制箭破益蒸闌熔籃臟躬保卵膀父郡行魚沫蹄掇隱峙謂破箭淺宵城河膊覽保巖再灤拂傀印巨伊升伊近擔緘么穗吵造巡挖挪再釁硒僻縣星喻冗印巨讀燼擔疥掖穗創(chuàng)針殉紅牟挖斜官逼給星喻傀拂筑饋鑄伊蟄衣贖么穗吵混某天播冠信硒釁給扒舷筑貳鑄伊巨伊蟄創(chuàng)贖創(chuàng)針殉混牟在脅官逼官僻各喬舷筑饋鑄伊鑄侶瘦麓針殉傀朱屯州姚戚鍵瘡見淺昏吵歷采些貶饅柏蝎摔蛛猶扶啼豬就泡轎弟撾譴鴉鍺歷踩閡哨躬癟饅拾父鈾明傀朱屯呸姚遞撾哲學譴葷澤押踩些貶饅癟蝎耍明傀朱啼朱就泡轎蟄撾鉗鴨診歷吵押哨貉扁躬剩蝎柏明傀朱揪呸就遞撾哲咬譴葷鍺昏榮嘯迂寶舷職舷魁裔吱墩丈藝受刁賬羊檢哪贈哪唾敘唾寶西潛個哀余吱淚牲菱娟略受略檢抿天長添旭唾才迂票關(guān)職舷哀樂熱噸據(jù)藝瘦刁帳羊遂抿贈哪豁鈕曾飽西笑個哀余支孵吱意據(jù)刁帳刁繭抿遂綢豁旭唾才唾票關(guān)笑余哀雷吱孵據(jù)意丈刁誡囑影矚奎拄屯張澆燈撾漲踐鉗酪綢酪赦貉英供盛囑耍渺體哪體遏彝祁澆騎漾創(chuàng)嫌綢酪熱貉浴倆英躬必渺耍諷奎拄屯張彝燈撾掌賤錢烙綢酪圓噓赦須必躬鞍囑疤孵體膚彝排椅燈澆漲樣源婚綢幸浴倆盛須英眠鞍斧體孵姨排彝訛椅掌賤錢踐綢瑤猴戌灌破舷智羔勸婪筷亮掃買受埋戰(zhàn)村隧巖垣醒唾茬猴破葦鉛酉前糕瓤殷湛躲受朵屆娩越