TWI351882B - - Google Patents

Download PDF

Info

Publication number
TWI351882B
TWI351882B TW96141499A TW96141499A TWI351882B TW I351882 B TWI351882 B TW I351882B TW 96141499 A TW96141499 A TW 96141499A TW 96141499 A TW96141499 A TW 96141499A TW I351882 B TWI351882 B TW I351882B
Authority
TW
Taiwan
Prior art keywords
low
module
image
performs
band phase
Prior art date
Application number
TW96141499A
Other languages
Chinese (zh)
Other versions
TW200922334A (en
Original Assignee
Univ Nat Chunghsing
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Univ Nat Chunghsing filed Critical Univ Nat Chunghsing
Priority to TW96141499A priority Critical patent/TW200922334A/en
Publication of TW200922334A publication Critical patent/TW200922334A/en
Application granted granted Critical
Publication of TWI351882B publication Critical patent/TWI351882B/zh

Links

Landscapes

  • Compression Or Coding Systems Of Tv Signals (AREA)

Description

1351882 九、發明說明: 【發明所屬之技術領域】 本發明為提供一種小波視訊編碼,採用離散小波轉換和移動估測方法 來達到壓縮效果,而達到運算複雜度和所需記憶體均較傳統方法來的少之 適用於小波視訊編碼之低複雜度頻域移動估測方法。1351882 IX. Description of the invention: [Technical field of the invention] The present invention provides a wavelet video coding, which uses a discrete wavelet transform and a motion estimation method to achieve a compression effect, and achieves computational complexity and required memory compared to conventional methods. Less of the low complexity frequency domain motion estimation method for wavelet video coding.

【先前技術】 按’伴隨著資訊的快速成長,影音的資料量也越來越魔大,應用的領 域也有巨大驗變,所轉像雜的重要性也就越來越突顯,在網路的蓬 勃發展之下,許多資料壓縮及資料傳輸技術也跟著被人們討論與研究,經 過多年的發展’壓織術也顯賴進步,_上已有技術非常成 熟的標準,其中最常見到的有ISO(Internationa 〇 rganization f〇r Standardization) 國際組織所制定的MP EG系列;另-侧是由ϊ τυ ( ! n t e r n ational Telecomm- nication Uni〇n )國際組織所制领H · 2 6 x系列,這兩組織後來共同成立-個團體為 WTCJoint Video Team)’共同制定了 η·26 4/ MPEG4 part 1〇此標準,為目前最新的影像壓縮標準主要 目的疋為了加速多媒體通信的發展’並促成電信、電腦、廣播電視三大網 路的融合’由於在壓縮品質或是壓驗果上,都有非常亮_表現,目前 網路上的電影❹舰大多靖PEG_4的方式來壓縮,如第三代行動 5 13518.82 通訊中的視訊服務也是。 數位視訊難技術發展已超過十年,賴也有小部份的演變,由不同 的研究理論根據大致可以歸納為下列幾種·· 第-種為以模組為基礎的,㈣像畫面上的文字 '背景及騎視為不 同物件,即物件模組化(〇bject mo d P 1 ·、 m〇cie 1 ! ng)’並制定以物 件為基礎的方絲驗畫面,以_最有效率_縮方式。在此—概 念中’將動祕像序列中的每-個場景,看成由不_視訊物件平面所组 #成。而同一物件的連續視訊物平面稱為視訊物件,其視訊物可以是動態影 像序列中的人或物或背景,即所輸人的動態影像,通過分析可將其分割成 數個視訊物件,而每個視訊物件再對其中所組成的視訊物件分別平面進行 編碼、儲存和傳送。至於解碼端再依不__來重組、峨或是替換所 需的視訊物件平面。 第二種方式為壓縮過程沒有做物件模組化,即壓縮方式是以一張張影 像晝面為主題進行的,將每張輸人的影像皆視為視訊訊號(s丨g n a丄 ^ b a S e d ) ’將影像訊號間多餘資料矛以刪除,因此,整張晝面上的所 有圖形、物件及背景,都採用相_壓縮編碼方式,雖然產生出來的資料 量相對的也較大,但她於物件模組化的高複雜度,且效率低,視訊訊號 擁有較簡單的架構及合理的編碼增益,所以物件模組化編碼方式大多應用 在影像合成或處理上。 為達到上述數位視訊壓縮技術,可透過使用轉補償系統和使用移動 補償滤波祕來完成,*飾猶綠又分為_散餘為基礎和以離 i 6 1351882 散小波為基礎;鶴補顧、波^純分料_域和鮮域飾補償編碼 _,其中,又則、波轉換的延展性(s c a i a b i } i t y )被廣範 的討論,基於小波影像的編解碼有2麵法,在頻域(麵為小波域)上 n —band)和空間域上(spa t i a 1:)。 而頻域執行編解碼的優點為: 1)編解碼端更容易實現空間的延展性(s p a t i a ability)。[Prior Art] According to the rapid growth of information, the amount of video and audio has become more and more powerful, and the field of application has also undergone tremendous changes. The importance of turning the image is becoming more and more prominent. Under the booming development, many data compression and data transmission technologies have been discussed and researched. After years of development, 'pressure weaving has also made progress. _There are already very mature standards, the most common ones are ISO. (Internationa 〇rganization f〇r Standardization) The MP EG series developed by international organizations; the other side is the H · 26 x series produced by ϊ υ ! ( ! ntern ational Telecomm- nication Uni〇n ) international organization. The organization was later established - a group for the WTCJoint Video Team) 'co-developed η·26 4/ MPEG4 part 1 〇 this standard, the main purpose of the latest image compression standard is to accelerate the development of multimedia communication' and promote telecommunications, computers The integration of the three major networks of radio and television 'because of the quality of the compression or the test results, there are very bright _ performance, the current movie on the Internet, most of the ❹ PEG_4 The way to compress, such as the third generation of action 5 13518.82 communication in the video service is also. The development of digital video hard technology has been more than ten years, and Lai has a small part of evolution. It can be summarized as follows based on different research theories. · The first type is based on modules, and (4) the text on the screen. 'Background and riding are regarded as different objects, that is, the object is modularized (〇bject mo d P 1 ·, m〇cie 1 ! ng)' and the object-based square silk inspection screen is formulated to be the most efficient the way. In this concept, each scene in the sequence of secret images is treated as a group of non-video objects. The continuous video object plane of the same object is called a video object, and the video object can be a person or object or a background in the motion image sequence, that is, a moving image of the input person, which can be divided into several video objects by analysis, and each The video objects then encode, store and transmit the video objects respectively formed in the plane. As for the decoding end, it is not necessary to reorganize, 峨 or replace the required video object plane. The second method is that the compression process is not modularized, that is, the compression method is based on the image of one image, and each input image is regarded as a video signal (s丨gna丄^ ba S Ed) 'The extra information between the video signals is removed. Therefore, all the graphics, objects and backgrounds on the entire surface are phase-compressed. Although the amount of data generated is relatively large, she The modularity of the object is high complexity and low efficiency. The video signal has a simple structure and reasonable coding gain. Therefore, the modular coding method of the object is mostly applied to image synthesis or processing. In order to achieve the above-mentioned digital video compression technology, it can be completed by using a rotation compensation system and using a mobile compensation filter. The *green is divided into _spread based and based on the scattered wavelet of i 6 1351882; Wave ^ pure material _ domain and fresh domain decoration compensation code _, which, in addition, the wave conversion ductility (scaiabi) ity is widely discussed, based on wavelet image codec has two-sided method, in the frequency domain (face is the wavelet domain) on n-band) and on the spatial domain (spa tia 1:). The advantages of performing codec in the frequency domain are as follows: 1) The codec is easier to implement spatial scalability (s p a t i a ability).

2) 在較低的frame rate下有較好的效能。 3) 在解析度較差的情況下有較好的效能。 有移=__換’再執行移動估測時,由於離散小波轉換 有移動變異特性(s h i f t v P r : ,2) Better performance at lower frame rates. 3) It has better performance in the case of poor resolution. When there is shift = __ change and then the motion estimation is performed, since the discrete wavelet transform has a mobile variation characteristic (s h i f t v P r : ,

Variant Pr〇pe 須利用屬於移動不變異的補償離散小波轉換(〇dwt ,而傳統的健帶她平移法所如_太大,以 博動估測 移法來估算,則f要料丨3 P自低頻帶相位平 异則篇要原來10倍錢體來儲存資料 到移動估測的運諸雜度。 "間接影響 7 丄乃1882 【發明内容】 故’發明人有鑑於上述缺失,乃搜集細資料,經由多方評估及考量 並Μ從事於此行業累積之多年經驗經由不斷試作及修改始設計出此 種採用離散、轉換和移祕财法來刺魏踩,*翻運算複雜度 和所需記細触傳财絲的处適鎌小舰鳩碼之倾雜度頻域 移動估測方法的發明專利者。Variant Pr〇pe must use the compensated discrete wavelet transform (〇dwt) which belongs to the mobile non-mutation, and the traditional kinetic band her translation method is too large, and is estimated by the pulsating estimation method, then f is expected to be 3P. Since the low-frequency phase is different, the original 10 times the amount of money to store the data to the mobile estimation is mixed. "Indirect impact 7 丄 is 1882 [Inventive content] Therefore, the inventor has collected the above-mentioned missing Detailed information, through multi-party assessment and consideration, and the years of experience accumulated in this industry, through continuous trial and modification, we have designed this kind of discrete, conversion and transfer financial method to stab the Wei step, the complexity and the required It is the inventor of the method of estimating the frequency domain of the small ship's weight.

本發私主要目_於:先將鱗㈣雜顿_,雜行移動估 測;但離散小波轉換有移動變異特性(shift variant D r 〇 P e r t y ),必翻用屬於移動不變異的補償離散小波轉換(⑽w Ή來執行移動估測’而低頻帶相位平移法(1gw f t )為猶離散小_換的频魏;但傳制低鮮她平移法所需 記憶體太大,蝴f 她规瓣,_相〇倍記 憶體來儲存資料,記憶體太大間接影響到移動估測的運算複雜度。本發明The main subject of this is _: firstly, the scale (4) is mixed with _, and the moving line is estimated; but the discrete wavelet transform has shift variability (shift variant D r 〇P erty ), which must be used to compensate for the discrete motion Wavelet transform ((10)w Ή to perform motion estimation' and low-band phase shift method (1gw ft) is the frequency of the _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The valve, _ phase 〇 memory to store data, the memory is too large and indirectly affects the computational complexity of the motion estimation. The present invention

簡化低頻帶相位平移法’只在第工階執行補償離散小波轉換也就是低頻帶 相位平移法,在較高的階層則使用離散小波轉換,以3階來估算,所需記 憶體梅的4倍;再_離散小波轉換之間不_之子母層關係和移 動向里(motion vectQr)共频念進行移動估測,藉此移 動估測可降低9 〇%的運算複雜度。 8 1351882 【實施方式】 為達成上述目的及功效,本發明所採用之技術手段及構造,茲繪圖就 本發明較佳實施例詳加說明其特徵與功能如下,俾利完全了解。 請參閱第-圖所示,係為本發明較佳實施例之方塊圖,由圖中可清楚 看出本發明係包括: —部份低相位平移模組E,將—參考圖像執行部份低鮮相位平 移; 一移動估測模組C,接收該部份低頻帶相位平移模組E傳輸之參考圖 像’並進行移動估測; 一移動補償模組D ’接收該移動估測模組㈣輸之參考圖像,並進行 移動補償; 離散小波職肋B,料敎—輸人娜錢補散小波轉換,Simplified low-band phase shifting method 'Compensates discrete wavelet transform only in the first step, that is, low-band phase shifting method. In the higher level, discrete wavelet transform is used, which is estimated by 3 steps, 4 times of the required memory plum. The motion estimation between the non-discrete wavelet transform and the motion vectQr co-frequency is performed, and the motion estimation can reduce the computational complexity by 9 〇%. 8 1351882 [Embodiment] In order to achieve the above object and effect, the technical means and structure of the present invention will be described in detail with reference to the preferred embodiments of the present invention. Referring to the first embodiment, which is a block diagram of a preferred embodiment of the present invention, it will be apparent from the drawings that the present invention includes: - a portion of the low phase translation module E, which will be - a reference image execution portion Low-precision phase shifting; a mobile estimation module C, receiving the reference image transmitted by the part of the low-band phase shifting module E and performing motion estimation; a mobile compensation module D' receiving the mobile estimation module (4) The reference image of the loss and the movement compensation; Discrete wavelet rib B, the material 敎 输 输 输 娜 钱 钱 钱 , ,

產生一目 ^ *圖像’該目前圓像與經該移動補償模組D處理過之參考圖像執 行相減運算並產生一殘餘訊號; 一編碼模組G,將該殘餘訊號轉為位元流η。 圖I二參閱第―、二圖所示’係為本發明較佳實施例之方塊圖與流程 圖,由圖中可清楚看出,藉由上述構件動作時,依照下列步 (1 m . J ' (11) 目_像執行離散小波轉換 e b c r e t e wave t transform» DWT); (12) 參考圖像執行部份低頻帶相位平移(P L s B); 9 1351882 (13)部份低頻帶相位平移係數内差; (1 4 )執行移動估測(motion estimati〇n); (1 5 )執行移動補償(mot ion compensa t ,· Λ L i ο Π (16)執行相減運算; (1 *7 )執行編瑪。 於上述步驟進行時’輸入影像A經由離散小波轉換模組8產生目前圖 screte wavele 像’目前圖像依序執行離散小波轉換(d i t t r a n s f o rm,DWT)以及由部份低頻帶相位平移模組£進 行部份低頻帶相位平移(P L S B )’將部份低頻帶相位平移係數内差由移 動估測模組C執行移動估測,續以移動補償模組D執行移動補償,且執行 相減運算產线餘喊’ _餘峨㈣碼额G執行編碼產生輸出位^ 請娜㈣為她⑽梅她输示意圖,由 圖中可清楚看出,將原始影像往右(χ m^-rr r ^ s > , )彺下(Y軸方向)和往右 p下(心軸再在y軸方向)移動,此時將有錄影像,再加 像,一共有杨影像,將這4張影像進行_小波轉換(dis„、;7 e wavelet tranSf0r 4個HL、4個LH和4個HH DWT),將產生4個LL、 〇)、⑻ 〇,〇)、hH(0,0):wh“0, 經過DWT產生4個sub —ba H L (〇,0)表示為原始影像 .HHM n^TT and;HL(l,0)、LH(1,0) 不為往右(X軸)方向移動後的影 1351882 像執行dw丁所產生的_sub—band,以此類推, 二 “,謝™ 一 Μ: 先將影像經過離散小波轉換,再執行移動估測 動變異特性“………t pr〇pe=:移 =移動不變異的補償離散小波轉換(。dwt)來執行移動估測,: :帶相位平移法U〇w band shi⑴為補伽散小波轉 =拖體加㈣她㈣她输,只顿階齡補償離散小 轉換也就是低鮮她平移法,在較高,以 3階來估所需記麵為原來的4倍;再利騰散小波轉換之間不同階 層之子母層關係和移動向量(m〇ti〇n vect〇r)共用觀念進 仃移動估測,藉此移動制可降低9 G %的運算複雜度。 淮’以上所述僅為本發明之較佳實施例而已,非因此即拘限本發明之 專利範圍,故舉凡運用本發明說明書及圖式内容所為之簡易修飾及等效結 構變化’均應同理包含於本發明之專利範圍内,合予陳明。 所述本發明之適用於小波視訊編碼之低複雜度頻域移動估測方 法於使用時’為確實能達職功效及目的,故本發明誠為-實祕優異之 發明’為符合發明專利之申請要件,爰依法提出申請,盼料早日賜准 本案’以轉發明人之辛苦發明,齡釣局審委有任何稽疑 ,請不吝來 1351882 函指示,發明人定當竭力配合,實感公便。Generating an image of the current image and performing a subtraction operation on the reference image processed by the motion compensation module D to generate a residual signal; an encoding module G, converting the residual signal into a bit stream η. Figure 2 is a block diagram and a flow chart of a preferred embodiment of the present invention. It can be clearly seen from the figure that when the above components are operated, the following steps are followed (1 m . J ' (11) _creature wave transform ebcrete wave t transform» DWT); (12) reference image performs partial low-band phase shift (PL s B); 9 1351882 (13) part of low-band phase shift coefficient Internal difference; (1 4) perform motion estimation (motion estimati〇n); (1 5) perform motion compensation (mot ion compensa t, · Λ L i ο Π (16) perform subtraction operation; (1 *7) Performing the editing. When the above steps are performed, 'the input image A generates the current graph scree wavelet image through the discrete wavelet transform module 8. The current image sequentially performs discrete wavelet transform (dittransfo rm, DWT) and is partially shifted by a low frequency band phase. The module performs part of the low-band phase shift (PLSB)' to perform the motion estimation by the motion estimation module C for the partial low-band phase shift coefficient internal difference, and continues to perform motion compensation with the motion compensation module D, and performs phase compensation. Subtracting the production line and shouting ' _ Yu Yu (four) code amount G Line coding produces the output bit ^ Please Na (4) for her (10) Mei she loses the schematic, as can be clearly seen from the figure, the original image to the right (χ m^-rr r ^ s >, ) 彺 (Y-axis direction) And move to the right p (the mandrel and then in the y-axis direction), at this time there will be recorded images, plus images, a total of Yang images, the four images are _ wavelet transform (dis„,; 7 e wavelet tranSf0r 4 HL, 4 LH and 4 HH DWT), will produce 4 LL, 〇), (8) 〇, 〇), hH (0, 0): wh "0, 4 sub-ba HL generated by DWT ( 〇, 0) is expressed as the original image. HHM n^TT and; HL(l,0), LH(1,0) is not moved to the right (X-axis) direction after the shadow 1351882 is generated by the implementation of dw Sub-band, and so on, two ", Xie TM one: first transform the image through discrete wavelet, and then perform the motion estimation dynamic mutation characteristics ".........t pr〇pe=: shift=moving non-mutation compensation discrete Wavelet transform (.dwt) to perform motion estimation, :: with phase shift method U〇w band shi(1) for complement gamma wavelet transform = drag body plus (four) she (four) she loses, only the age level compensation discrete small conversion is low Fresh her translation method, in comparison High, the 3rd order is used to estimate the required 4 times of the original; the different layers of the child-level relationship and the moving vector (m〇ti〇n vect〇r) between the different wavelet transforms are shared. By measuring, the mobile system can reduce the computational complexity of 9 G %. The above description is only the preferred embodiment of the present invention, and thus the scope of the present invention is not limited thereto, so that the simple modification and the equivalent structural change of the present invention and the contents of the drawings are the same. It is included in the scope of the patent of the present invention and is combined with Chen Ming. The low-complexity frequency-domain mobile estimation method applicable to the wavelet video coding of the present invention is used for the purpose of achieving the function and purpose, so the present invention is a true invention and is excellent in conformity with the invention patent. Apply for the requirements, and apply in accordance with the law, I hope that the case will be granted as soon as possible. In order to transfer the inventor's hard work, the trial committee of the age-fishing bureau has any doubts. Please do not hesitate to give instructions to the 1351882 letter. The inventor will try his best to cooperate and feel polite.

12 13518.82 【圖式簡單說明】 第一圖係為本發明較佳實施例之方塊圖。 第二圖係為本發明較佳實施例之流程圖。 第三圖係為本發明之部份低頻帶相位平移法示意圖。 【主要元件符號說明】12 13518.82 BRIEF DESCRIPTION OF THE DRAWINGS The first drawing is a block diagram of a preferred embodiment of the invention. The second drawing is a flow chart of a preferred embodiment of the present invention. The third figure is a schematic diagram of a part of the low frequency band phase shifting method of the present invention. [Main component symbol description]

輸入影像........AInput image........A

離散小波轉換模組····BDiscrete Wavelet Transform Module····B

移動估測模組......CMobile estimation module...C

移動補償模組......DMotion compensation module...D

部份低頻帶相位平移模組·EPart of the low-band phase shifting module·E

反離散小波轉換模組···FAnti-Discrete Wavelet Transform Module···F

編碼模組........G 位元流.........Η 13Encoding module........G bit stream.........Η 13

Claims (1)

1351882 _ 時少〇曰修正替換頁100年·/月外修正替換頁 十、申請專利範圍: 1、一種適用於小波視訊編碼之低複雜度頻域移動估測方法,包括有: 一部份低頻帶相位平移模組,將一參考影像執行部份低頻帶相位平移; 一移動估測模組’接收該部份低頻帶相位平移模組傳輸之參考影像並 進行移動估測; 償 •移動補償餘,接⑽移祕繼組雜之參考树,錢行移動補 -離散小波轉換模組,將外部之一輸入影像進行離散小_換產生— 目前圖像’該目前圖像與經該移動補償模組處理過之參考_執行相減 運算並產生一殘餘訊號; 一編碼模組,將該殘餘訊號轉為位元流。 4碥瑪之低複雜度頻 2、如申請專利範圍第i項所述之一種適用於小波視〗 域移動估測方法,該方法為依照下列步驟進行: 一、 目前影像執行離散小波轉換; 二、 參考職執行部份低解相位平移; 三、將部份低頻帶相位平移係數内插執行移動估測; 四、 執行軸娜純行_產錢餘訊號; 五、 將殘制號執行編韻生輸φ位元流。 141351882 _ 〇曰 〇曰 〇曰 替换 替换 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 The band phase shifting module performs a partial low-band phase shift on a reference image; a mobile estimation module receives the reference image transmitted by the part of the low-band phase shifting module and performs motion estimation; , (10) to move the secret reference group of the reference tree, the money line mobile complement-discrete wavelet transform module, the external input image is discretely small _ change generated - the current image 'the current image and the mobile compensation mode The group processed reference _ performs a subtraction operation and generates a residual signal; an encoding module converts the residual signal into a bit stream. 4 碥 之 低 复杂 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 2、 The reference part performs low-resolution phase shifting. 3. The partial low-band phase shift coefficient is interpolated to perform the motion estimation. 4. The execution axis Na pure line _ the production residual signal; 5. The residual number is executed. The φ bit stream is generated. 14
TW96141499A 2007-11-02 2007-11-02 Low complexity frequency field motion estimation method adapted to wavelet video coding TW200922334A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW96141499A TW200922334A (en) 2007-11-02 2007-11-02 Low complexity frequency field motion estimation method adapted to wavelet video coding

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW96141499A TW200922334A (en) 2007-11-02 2007-11-02 Low complexity frequency field motion estimation method adapted to wavelet video coding

Publications (2)

Publication Number Publication Date
TW200922334A TW200922334A (en) 2009-05-16
TWI351882B true TWI351882B (en) 2011-11-01

Family

ID=44728143

Family Applications (1)

Application Number Title Priority Date Filing Date
TW96141499A TW200922334A (en) 2007-11-02 2007-11-02 Low complexity frequency field motion estimation method adapted to wavelet video coding

Country Status (1)

Country Link
TW (1) TW200922334A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8452605B2 (en) 2006-10-25 2013-05-28 Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. Apparatus and method for generating audio subband values and apparatus and method for generating time-domain audio samples
TWI568249B (en) * 2011-11-03 2017-01-21 太陽專利信託 Efficient rounding for deblocking

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2982983A1 (en) * 2011-11-22 2013-05-24 Thomson Licensing METHOD FOR ENCODING AND RECONSTRUCTING A BLOCK OF PIXELS AND CORRESPONDING DEVICES

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8452605B2 (en) 2006-10-25 2013-05-28 Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. Apparatus and method for generating audio subband values and apparatus and method for generating time-domain audio samples
TWI568249B (en) * 2011-11-03 2017-01-21 太陽專利信託 Efficient rounding for deblocking

Also Published As

Publication number Publication date
TW200922334A (en) 2009-05-16

Similar Documents

Publication Publication Date Title
Jiang et al. An end-to-end compression framework based on convolutional neural networks
Jin et al. A flexible deep CNN framework for image restoration
US20060120448A1 (en) Method and apparatus for encoding/decoding multi-layer video using DCT upsampling
Zhou et al. Scalable compression of stream cipher encrypted images through context-adaptive sampling
JP2008541653A (en) Multi-layer based video encoding method, decoding method, video encoder and video decoder using smoothing prediction
WO2008145039A1 (en) Methods, systems and devices for generating upsample filter and downsample filter and for performing encoding
CN105408935A (en) Upsampling and signal enhancement
KR20210017185A (en) Method and apparatus for removing compressed poisson noise of an image based on deep neural network
CN102917226B (en) Intra-frame video coding method based on self-adaption downsampling and interpolation
TWI351882B (en)
CN1255988A (en) Down scaling of images
CN102075766B (en) Video coding and decoding methods and devices, and video coding and decoding system
Zhao et al. Lighter but efficient bit-depth expansion network
CN101083769A (en) Coder/decoder and coding/decoding method for processing N bit video data
CN1809169A (en) Method and apparatus for encoding pictures without loss of DC components
Tan et al. Image compression algorithms based on super-resolution reconstruction technology
JP7125559B2 (en) Video stream adaptive filtering for bitrate reduction
Mehrseresht et al. Spatially continuous orientation adaptive discrete packet wavelet decomposition for image compression
Qi et al. CNN-Based Post-Processing Filter for Video Compression with Multi-Scale Feature Representation
CN103402090A (en) Improving method based on Huffman lossless compression algorithm
Huang et al. Mode Information Guided CNN for Quality Enhancement of Screen Content Coding
Jia et al. Deep convolutional network based image quality enhancement for low bit rate image compression
CN101146227A (en) Build-in gradual flexible 3D wavelet video coding algorithm
Gao et al. Extremely Low Bit-rate Image Compression via Invertible Image Generation
CN115358954B (en) Attention-guided feature compression method

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees