以作者查詢圖書館館藏 、以作者查詢臺灣博碩士 、以作者查詢全國書目 、勘誤回報 、線上人數:21 、訪客IP:3.149.25.162
姓名 楊智強(Zhi-Qiang Yang) 查詢紙本館藏 畢業系所 通訊工程學系 論文名稱 基於CNN應用於VVC碼率控制之研究
(CNN-Based Approach to Rate Control in VVC)相關論文 檔案 [Endnote RIS 格式] [Bibtex 格式] [相關文章] [文章引用] [完整記錄] [館藏目錄] 至系統瀏覽論文 (2027-1-22以後開放) 摘要(中) 在現今的社會,我們對於解析度的要求越來越高,為了因應我們所需高解析度的影像,多功能影像編碼(VVC)能比上一代的視訊編碼(HEVC)高出兩倍的壓縮率,這是因為在HEVC的壓縮技術中,使用了編碼單元、預測單元、轉換單元以及量化等方式。在網路傳輸方面,碼率控制是為了使傳輸的影像在特定的通道容量下有較低的失真量以及較好的效能,本論文採用R-λ model控制在low-delay 配置下CTU級的比特率,對於畫面間預測,發現部分CTU的位元錯誤率和像素域的變異數以及運動向量的變異數呈正相關,在R-λ model下的碼率控制算法非常依賴λ值的精確度,達成精確的目標比特率,在本論文中,為了能使畫面間預測有更精確的碼率控制方法,利用統計決定像素域變異數以及運動向量變異數特定閥值,若大於閥值,則引用卷積神經網路來預測編碼樹單元的參數,實驗結果表明,基於卷積神經網路的方法,相較於多功能影像編碼中VTM 20.0的碼率控制方法的部分,在位元錯誤率方面降低了0.122%。 摘要(英) In today′s society, our demand for resolution is increasing. To meet the requirements for high-resolution images, Versatile Video Coding (VVC) can achieve a compression rate twice as high as the previous generation video coding standard, High Efficiency Video Coding (HEVC). This improvement is attributed to the use of coding units, prediction units, transform units, and quantization in HEVC′s compression techniques.
In terms of network transmission, bitrate control aims to achieve lower distortion and better performance of transmitted images within a specific channel capacity. This paper adopts an R-λ model to control the bit rate at the Coding Tree Unit (CTU) level under low-delay configuration. For inter-frame prediction, it is observed that the bit error rate of some CTUs is positively correlated with pixel domain variance and motion vector variance. The bitrate control algorithm under the R-λ model heavily relies on the accuracy of the λ value to achieve precise target bit rates.
In this paper, for more accurate bitrate control in inter-frame prediction, a statistical approach is used to determine specific thresholds for pixel domain variance and motion vector variance. If these variances exceed the threshold, a Convolutional Neural Network (CNN) is invoked to predict the parameters of the coding tree unit. Experimental results demonstrate that the CNN-based approach, compared to the bitrate control method in VTM 20.0 of VVC, reduces the bit error rate by 0.122%.關鍵字(中) ★ VVC
★ 畫面間預測
★ 碼率控制
★ RDO關鍵字(英) ★ VVC
★ inter prediction
★ rate control
★ RDO論文目次 第一章 緒論 1
1.1多功能影像編碼(Versatile Video Coding, VVC) 1
1.2 VVC編碼架構 1
1.2.1編碼單元(Coding Unit)、預測單元(Prediction Unit)、轉換單元(Transform Unit) 3
1.2.2量化參數(Quantization Parameter) 5
1.2.3 H.265/HEVC 和 H.266/VVC 差異 6
1.3支持向量機(Support Vector Machine)介紹 7
1.4深度學習介紹 8
1.4.1類神經網路 9
1.4.2深度學習 9
1.5研究動機 11
1.6論文架構 12
第二章 相關文獻 12
2.1碼率控制(Rate Control) 12
2.2 A convolutional neural network-based approach to rate control in HEVC intra coding 17
2.2.1 整體系統架構介紹 17
2.2.2 實驗方法 18
2.2.3實驗結果 22
2.3 SVM應用於HEVC編碼單元(CU)快速深度決策演算法 23
2.3.1支持向量機(Support Vector Machine,SVM)特徵選取 23
2.3.2 快速深度決策演算法 26
2.3.3 模型訓練&量化參數的選擇 27
2.3.4 模型樣本數量縮減 31
2.3.5 實驗結果 32
2.4 基於SVM與CNN的碼率控制方法應用於畫面內編碼 33
2.4.1整體系統架構 34
2.4.2 CNN模型架構及訓練 35
2.4.3 CNN模型效能測試 38
2.4.4 CNN模型在HEVC編碼軟體中的應用 38
2.4.5實驗結果與分析 39
2.4.6整體實驗結果 40
第三章 基於CNN與SVM的碼率控制方法應用於VVC畫面內預測編碼 43
3.1系統架構 43
3.1.1CNN模型 43
3.1.2HEVC/VVC畫面內預測編碼比較 44
3.1.3SVM分類效能分析 50
3.2應用於VVC畫面內預測編碼之實驗結果與分析 52
3.2.1相關函數定義 52
3.2.2基於SVM/CNN應用於VVC/HEVC碼率控制實驗結果之比較 53
第四章 基於CNN的碼率控制方法應用於VVC畫面間預測編碼 57
4.1系統架構 57
4.2CNN模型 58
4.2.1像素域變異數(pixel variance)與運動向量變異數(MV variance) 58
4.2.2CNN模型訓練與效能測試 62
4.3基於CNN應用於VVC碼率控制實驗結果之比較與分析 63
第五章 參考文獻 70參考文獻 [1] Ye Li , Bin Li, Dong Liu, Zhibo Chen,“RATE CONTROL FOR VERSATILE VIDEO CODING,”in 2020 IEEE International Conference on Image Processing (ICIP)
[2] Y. Li, B. Li, D. Liu and Z. Chen, "A convolutional neural network-based approach to rate control in HEVC intra coding," 2017 IEEE Visual Communications and Image Processing (VCIP), 2017, pp. 1-4, doi: 10.1109/VCIP.2017.8305050.
[3] S.J. Cai, “Reduction of computation complexity for HEVC intra prediction with support vector machine,” National Central University, Master Thesis, Jun 2017.
[4]D.M. Lin, “SVM/CNN-Based Approach to Rate Control in HEVC Intra Coding
,“National Central University, Master Thesis, Jun 2021
[5] B. Li, H. Li, L. Li and J. Zhang, “λ domain rate control algorithm for high efficiency video coding”, IEEE Transactions on Image Processing, vol. 23, no. 9, pp. 3841-3854, Sept 2014.
[6] Benjamin Bross, Ye-Kui Wang, Yan Ye, Shan Liu , Jianle Chen , Gary J. Sullivan , Jens-Rainer Ohm ,“Overview of the Versatile Video Coding (VVC) Standard and Its Applications,” IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 31, NO. 10, OCTOBER 2021
[7] Jianle Chen, “Algorithm description for Versatile Video Coding and Test Model 11 (VTM 11),” Document: JVET-T2002-v2 ,Joint Video Experts Team (JVET)
of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29
20th Meeting, by teleconference, 7 – 16 October 2020
[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] H. Choi, J. Nam, J. Yoo, D. Sim, and I. Bajic, “Rate control based on unified RQ model for HEVC,” ITU-T SG16 Contribution, JCTVCH0213, pp. 1–13, 2012.
[10] ISO/IEC JTC1/SC29/WG11, “Rate control for VVC,” Doc. JVET-K0390, Ljubljana, July 2018.
[11]G. Xu, S. Jin, KaichenTang, Z. Zhang and Y. Zhou, "LCU-level RateDistortion Opti-mization for Versatile Video Coding," 2022 IEEE International Symposium on Circuits and Systems (ISCAS), Austin, TX, USA, 2022, pp. 3023-3027, doi: 10.1109/ISCAS48785.2022.9937464.
[12] Y. Li, H. Jia, P. Ma, C. Zhu, X. Xie and W. Gao, "Inter-dependent ratedistortion modeling for video coding and its application to rate control," 2014 IEEE International Con-ference on Multimedia and Expo (ICME), 2014, pp. 1-6, doi: 10.1109/ICME.2014.6890338.
[13] Y. Mao, M. Wang, S. Wang and S. Kwong, "High Efficiency Rate Control for Versatile Video Coding Based on Composite Cauchy Distribution," in IEEE Transactions on Cir-cuits and Systems for Video Technology, vol. 32, no. 4, pp. 2371-2384, April 2022, doi: 10.1109/TCSVT.2021.3093315.指導教授 林銀議(Yin-Yi Lin) 審核日期 2024-1-23 推文 facebook plurk twitter funp google live udn HD myshare reddit netvibes friend youpush delicious baidu 網路書籤 Google bookmarks del.icio.us hemidemi myshare