跳到主要內容

簡易檢索 / 詳目顯示

回結果列表
研究生： 楊得弘
De-Hong Yang
論文名稱： SVM-CNN應用於HEVC畫面間編碼樹單元切割
SVM/CNN-based CTU Partition for HEVC Inter Prediction
指導教授： 林銀議
Yin-Yi Lin
口試委員:
學位類別： 碩士
Master
系所名稱： 資訊電機學院 - 通訊工程學系
Department of Communication Engineering
論文出版年： 2021
畢業學年度： 109
語文別： 中文
論文頁數： 130
中文關鍵詞： 高效率視頻編碼 、畫面間預測 、編碼單元 、快速深度決策 、支持向量機 、卷積神經網路 、碼率失真最佳化
外文關鍵詞： High efficiency video coding, inter picture prediction, coding unit, fast depth decision, support vector machine, convolutional neural network, rate distortion optimization
相關次數： 點閱：22  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 隨著科技日新月異及人們嚮往高解析度所帶來的影像品質，因此高解析度的影像產品也跟著與日俱增，以至於為了能夠有效壓縮高解析度影像膨大資料量，HEVC(High Efficiency Video Coding)採用了許多更新穎的技術來降低位元率，例如:畫面內預測、畫面間預測、碼率失真最佳化等等，但也同時造成了編碼計算複雜度提升。而本論文利用近幾年蓬勃發展起來的深度學習與機器學習中的卷積神經網路 CNN ( Convolutional Neural Network ) 及 支持向量機 SVM ( Support Vector Machine ),將其應用於HEVC編碼單元階段的決策。本論文在編碼一開始時先使用SVM來對編碼單元深度以及預測單元模式做分類，編碼單元以畫面間預測的移動向量值的資訊、合併模式的CBF、鄰近區塊深度資訊作為特徵(Feature)將一個CTU分類成只處理深度0、深度0~1、深度0~2、深度0~3四種類別，再利用卷積神經網路分層向下細分。藉由原本HEVC遞迴運算處理編碼單元的方式，在特定深度的編碼提前終止後續的編碼計算，以此節省後續深度所需計算時間達成編碼端縮減時間。最終實驗結果顯示，與HEVC相比，整體平均BDBR上升0.89%的情況下，編碼時間大約可以節省44%。

    With the rapid development of technology and the desire for high resolution, the image products with high resolution are increasing. In order to effectively compress the volume of high-resolution image expansion data, HEVC (high efficiency video coding) adopts many more novel technologies to reduce bit rate, such as: in-picture prediction, inter picture prediction, and so on The optimization of bit rate distortion and so on also leads to the increase of the complexity of coding calculation. In this paper, we use convolutional neural network (CNN) and support vector machine (SVM) in deep learning and machine learning to make decision in hevc coding unit stage. At the beginning of coding, this paper uses SVM to classify coding unit depth and prediction unit mode. Coding unit classifies a CTU into four categories: depth 0, depth 0 ~ 1, depth 0 ~ 2 and depth 0 ~ 3 based on the information of motion vector value predicted between pictures, CBF of merging mode and depth information of adjacent blocks, Then the convolution neural network is used to subdivide downward. By using the original hevc recursive operation to process the coding unit, the subsequent coding calculation is terminated in advance at a specific depth of coding, so as to save the calculation time of the subsequent depth and reduce the coding time. The final experimental results show that compared with HEVC, when the overall average bdbr increases by 1.07%, the encoding time can be saved by about 45%.

    論文摘要 I Abstract II 誌謝 IV 章節目錄 V 附圖索引 VIII 附表索引 XII 第1章 緒論 1 1.1高效率視訊編碼(HEVC)標準介紹 1 1.2高效率視訊編碼架構介紹 2 1.2.1編碼單元(Coding Unit) 3 1.2.2預測單元(Prediction Unit) 5 1.2.3轉換單元(Transform Unit) 6 1.2.4碼率失真代價函數(RD cost) 9 1.2.5 HEVC架構(Configuration) 10 1.3研究動機及目的 13 1.4論文架構 14 第2章 畫面間預測模式及支持向量機與卷積神經網路介紹 15 2.1 畫面間預測介紹(Inter Prediction) 15 2.1.1合併模式決策介紹(Merge Mode Decision) 15 2.1.2畫面間模式決策介紹(Inter Mode Decision) 18 2.2支持向量機(Support Vector Machine) 25 2.3 深度學習(Deep Learning) 28 2.3.1 類神經網路(Neural Network) 28 2.3.2 卷積神經網路(Convolutional Neural Network) 30 第3章 相關文獻回顧 35 3.1 利用SVM減少CU編碼複雜度文獻回顧 35 3.1.1支持向量機應用於HEVC畫面間編碼單元快速決策演算法 35 1. 移動向量變異數(Motion Vector Variance) 38 2. Coded Block Flag (CBF) 42 3. 鄰近編碼單元深度資訊 (Neighboring CU) 43 1. 訓練樣本(Training) 45 3. 效能分析及討論 48 3.2 利用CNN減少CU編碼複雜度文獻回顧 53 3.2.1 Fast CU Depth Decision for HEVC Using Neural Networks 53 3.3 利用結合SVM與CNN減少CU編碼複雜度文獻回顧 58 3.2.2 Computation Reduction of HEVC Intra Prediction using combined SVM and CNN 58 第4章 結合SVM與CNN應用於畫面間編碼區塊快速深度決策演算法 61 4.1 快速編碼單元決策演算法 62 4.1.1 階段一 64×64卷積神經網路 63 4.1.2 階段二 32×32卷積神經網路 63 4.1.3 階段三 16×16卷積神經網路 64 4.2整體系統流程 67 4.2.1 前處理階段 68 4.2.2 訓練階段 71 4.2.3 測試階段 78 4.3 演算法性能比較 86 4.3.1 CNN Level 1 性能比較 86 4.3.2 CNN Level 1+2 性能比較 88 4.3.3 CNN Level 1+2+3 性能比較 90 4.3.4 效能分析 93 第5章 原始圖像與殘差圖像模型性能分析 99 5.1 編碼單元分割與殘餘值(residual)圖片分析 99 5.1.1變更卷積神經網路輸入圖像資訊 100 5.1.2 性能比較 102 5.2 總性能比較 104 第六章 結論與未來展望 107 參考文獻 108

    [1] I. E. G. Richardson, H.264 and MPEG-4 Video Compression: Video Coding for Next-generation Multimedia. Aberdeen, U.K.: John Wiley & Sons, 2003.
    [2] “Generic coding of moving pictures and associated audio information,” ISO/IEC 13818-2: Video (MPEG-2), May 1996.
    [3] “Coding of audio-visual objects - Part 2: Visual,” in ISO/IEC 14496-2 (MPEG-4 Visual Version 1), Apr. 1999.
    [4] “Video coding for low bit rate communication, version 1,” ITU-T recommendation H.263, 1995.
    [5] Gary J. Sullivan, Jens-Rainer Ohm, Woo-Jin Han and Thomas Wiegand, “Overview of the high efficiency video coding (HEVC) Standard,” in Proc. IEEE Transactions on circuits and systems for video technology, vol. 22, no. 12, pp. 1649-1668, Dec. 2012.
    [6] P. Helle, S. Oudin, B. Bross, D. Marpe, M. O. Bici, K. Ugur, J. Jung, G. Clare, and T. Wiegand, “Block merging for quadtree-based partitioning in HEVC,” in Proc. IEEE Transactions on circuits and systems for video technology, vol. 22, no. 12, pp. 1720-1731, Dec. 2012.
    [7] L. Zhao, X. Guo, S. Lei, S. Ma and D. Zhao, “Simplified AMVP for high efficiency video coding,” in Proc. IEEE ICIP, pp. 1-4, 27-30 Nov. 2012.
    [8] J. L. Lin, Y. W. Chen, Y. W. Huang, and S. M. Lei, “Motion vector coding in the HEVC standard,” in Proc. IEEE Journal of Selected Topics in Signal Processing, vol. 7, no. 6, pp. 957-968, 3 July 2013.
    [9] Y. Ismail and S. El-etriby, “Fast diamond search algorithm for real time video coding,” in Proc. IEEE ICNC, pp. 729-733, Feb. 2012.
    [10] J.K. Liu, “Efficient HEVC inter prediction using SVM,” Department of Communication Engineering National Central University, Taiwan 32054, R.O.C., Jan 2019.
    [11] S.J Cai, “Reduction of computation complexity for HEVC intra prediction with support vector machine,” National Central University, Master Thesis, Jun 2017.
    [12] Tianyi Li,Mai Xu,Xin Deng, “ A deep convolutional neural network approach for complexity reduction on intra-mode HEVC”, 2017 IEEE International Conference on Multimedia and Expo (ICME).
    [13] K. Kim and W. W. Ro, "Fast CU depth decision for HEVC using neural networks", IEEE Trans. Circuits Syst. Video Technol., vol. 29, no. 5, pp. 1462-1473, May 2019.
    [14] Y. Li, Z. Liu, X. Ji and D. Wang, "CNN Based CU Partition Mode Decision Algorithm for HEVC Inter Coding", 2018 25th IEEE International Conference on Image Processing (ICIP), pp. 993-997, 2018.
    [15] Tzong-Dar Wu,Yuting Yen,J. H. Wang,R. J. Huang;Hung-Wei Lee and Hsuan-Fu Wang, "Automatic Target Recognition in SAR Images Based on a Combination of CNN and SVM",2020 International Workshop on Electromagnetics: Applications and Student Innovation Competition (iWEM).
    [16] D. B. Desai and S. N. Kavitha, "Face anti-spoofing technique using CNN and SVM", Proc. Int. Conf. Intell. Comput. Control Syst. (ICCS), pp. 37-41, May 2019.
    [17] D. U. N. Qomariah, H. Tjandrasa and C. Fatichah, "Classification of Diabetic Retinopathy and Normal Retinal Images using CNN and SVM", 2019 12th International Conference on Information & Communication Technology and System (ICTS), pp. 152-157, 2019.
    [18] Jie-Jay Wang, Yin yi Lin ,“Computation Reduction of HEVC Intra Prediction using combined SVM and CNN”, National Central University, Master Thesis, Jan 2020.
    [19] Hao-Chiun Wang, Yin yi Lin ,“CNN-based CU Partition for HEVC Intra Prediction”, National Central University, Master Thesis, July 2020.

    QR CODE
    :::