(3.235.191.87) 您好!臺灣時間:2021/05/14 20:27
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

: 
twitterline
研究生:黃盟淵
研究生(外文):meng–yuan huang
論文名稱:突顯視覺關注物件之2D/3D 轉換系統
論文名稱(外文):A 2D-to-3D Convert System with Visual Saliency Preservation
指導教授:張雲龍張雲龍引用關係陳洳瑾
指導教授(外文):Dr.Weng-Long ChangDr.Ju-Chin Chen
學位類別:碩士
校院名稱:國立高雄應用科技大學
系所名稱:資訊工程系
學門:教育學門
學類:專業科目教育學類
論文種類:學術論文
論文出版年:101
畢業學年度:100
語文別:中文
論文頁數:50
中文關鍵詞:深度圖顯著圖
外文關鍵詞:Depth mapSaliency map
相關次數:
  • 被引用被引用:1
  • 點閱點閱:965
  • 評分評分:
  • 下載下載:63
  • 收藏至我的研究室書目清單書目收藏:0
論文提出一個基於視覺注意突顯主體的2D/3D轉換,這種轉換是以2D影像加上以2D影像中各種特徵推估出的深度圖當輸入,最後依據深度影像渲染(Depth Image-Based Rendering,DIBR)轉換成立體影像對,大腦接收到影像後,會將此差異自然融合成兩個略有差異的影像,就變成一個帶有深度訊息(具有遠近線索)的視覺影像,即產生立體感。

在深度資訊估測上,以往深度估測只考慮到整個場景的深度感,人類對於產生的立體影像感官上主體不夠明顯,所以我們提出的方式是利用視覺注意的顯著圖和物體切割整合在深度計算上;除了保留了原有的層次感,更加強了視覺關注上的物件。

實驗將影像依照主體性的多寡與背景複雜程度分六類,共測試了九十張影像,並且使用本論文提出方式與其它兩種方式(視覺關注與MST切割法)做出的立體影像進行投票比較,實驗結果顯示在單一主體且背景單調的影像可以有最好的立體效果。
In this article, we propose a 2D to 3D video conversion system for outstanding lead based on visual attention. Original 2D image and estimate depth information among 2D image are treated as input image. Finally, the 3D stereo video can be readily generated by using a Depth Image-Based Rendering (DIBR) algorithm. Different views of the image can be synthesized by brain.

Different from the past that only consider entire conditions depth order among estimate depth map. Human no highly visually salient on 3D image. We propose a 2D/3D conversion system that integrate saliency map and object-segment on estimate depth map. Not only retain the original layering, but also to strengthen the visual attention on the object.

According to the amount of subjectivity and background complexity degree divided into six kinds for the experimental image with ninety. We compare we propose system, and the other two ways (visual attention method and MST segmentation method) by Vote. The experimental results show that the proposed algorithm is highly promising for 2D-3D convert system on single lead and monotonous background.
摘 要
ABSTRACT
致謝
目錄
表目錄
圖目錄
一、 緒論
1.1 研究動機
1.2 系統架構
1.3 論文架構
二、 相關文獻探討
2.1 深度估測
2.2 DIBR演算法
三、 基於線性透視與大氣透視法之深度圖之生成
3.1 判定拍攝方向
3.1.1 以線性透視當深度線索
3.1.2 以大氣透視當深度線索
3.1.3 合併兩深度線索
3.2 以最小生成展開樹(MST)之物件切割
3.3 生成初步深度圖
四、 視覺注意的提取及新視角的產生
4.1 顯著圖的產生與分割
4.2 深度圖與顯著圖的合成
4.3 立體影像對產生
五、 實驗結果與分析
5.1 測試資料收集與建立
5.2 實驗測試結果與分析
六、 結論與未來展望
參考文獻
[1] Itti, Koch, "A Saliency-Based Search Mechanism For Overt And Covert Shifts Of Visual Attention", Vision Res., pp. 1489-1506, 1999
[2] 黃怡菁, 黃乙白, 謝漢萍, "3D立體顯示技術", 科學發展, 451期, 2010
[3] 許精益、黃乙白, "3D立體顯示技術之發展與研究", 光學工程 第九十八期, 2010
[4] 賴文能、陳韋志, "淺談 2D 至 3D 視訊轉換技術", 影像與識別, Vol.16 No.2, 2010
[5] Foundations of Cyclopean Perception, Bela Julesz, The MIT Press.
[6] Ianir Ideses, Leonid Yaroslavsky, Barak Fishbain., "Depth Map Manipulation For 3D Visualization", 3DTV Conference: The True Vision - Capture, Transmission and Display of 3D Video, pp.337-340, 2008
[7] Yu-Lin Chang, Chih-Ying Fang, Li-Fu Ding, Shao-Yi Chen, and Liang-Gee Chen, 2007, "Depth Map Generation for 2D-to-3D Conversion By Short-Term Motion Assisted Color Segmentation", Multimedia and Expo, 2007 IEEE International Conference on,pp.1958-1961, 2007
[8] Chao-Chung Cheng, Chung-Te Li, and Liang-Gee Chen, Fellow, "A Novel 2D-to-3D Conversion System Using Edge Information", Consumer Electronics, pp.1739-1745, 2010
[9] Ching-Lung Su, Kang-Ning Pang, Tse-Min Chen, Guo-Syuan Wu, Chia-Ling Chiang, Hang-Rnei Wen, Lung-Sheng Huang, Ya-Hsin Hsueh, Shau-Yin Tseng, "A Real-time Full-HD 2D-to-3D Conversion System Using Multicore Technology", Multimedia and Ubiquitous Engineering, pp.273-276, 2011
[10] Guo-Shiang Lin, Cheng-Ying Yeh, Wei-Chih Chen, and Wen-Nung Lie, "A 2D to 3D Conversion Scheme Based On Depth Cues Analysis For MPEG Videos", Multimedia and Expo (ICME), pp.1141-1145, 2010
[11] Huabing Wang, Yingyun Yang, La Zhang, Yansi Yang, Bo Liu, "2D-to-3D Conversion Based On Depth-from-Motion", Mechatronic Science, Electric Engineering and Computer (MEC), pp.1892-1895, 2011
[12] Jiahong Zhang, You Yang, Qionghai Dai, "A Novel 2D-to-3D Scheme By Visual Attention And Occlusion Analysis", 3DTV Conference, pp.1-4,16-18, 2011
[13] C. Fehn, "A 3D-TV Approach Using Depth-Image-Based Rendering (DIBR)", in Proceedings of 3rd IASTED Conference on Visualization, Imaging, and Image Processing, Benalm´adena, Spain, pp. 482–487, 2003
[14] Liang Zhang, Wa James Tam, "Stereoscopic Image Generation Based On Depth Images For 3D TV", Broadcasting, IEEE Transactions on, pp.191-199, 2005
[15] Wan-Yu Chen, Yu-Lin Chang, Shyh-Feng Lin, Li-Fu Ding, and Liang-Gee Chen, "Efficient Depth Image Based Rendering With Edge Dependent Depth Filter And Interpolation", ICME, pp.1314-1317, 2005
[16] Kwan-Jung Oh, Sehoon Yea, Yo-Sung Ho, "Hole-Filling Method Using Depth Based In-Painting For View Synthesis In Free Viewpoint Television (FTV) and 3D Video", PCS, pp.1-4, 2009
[17] Hyun-Woong Cho, Soon-Wook Chungt, Moon-Kyu Song, and Woo-Jin Song, "Depth-Image-Based 3D Rendering With Edge Dependent Preprocessing", Circuits and Systems (MWSCAS), pp.1-4,7-10, 2011
[18] 劉然, "一種用於DIBR"的去隔行算法"
[19] 劉楷哲, "基於3D顯示器格式之即時3D內容合成技術", 影像與識別, Vol.16 No.2, 2010
[20] http://www.scholarpedia.org/article/Visual_salience
[21] Ming-Ming Cheng, Guo-Xin Zhang Niloy J. Mitra Xiaolei Huang Shi-Min Hu, "Global Contrast based Salient Region Detection", Computer Vision and Pattern Recognition (CVPR), pp.409-416,20-25, 2011
[22] P. Felzenszwalb and D. Huttenlocher., 2004, "Efficient Graph-Based Image Segmentation" IJCV, 59(2):167–181, 2004
[23] C. Rother, V. Kolmogorov, and A. Blake., "Grabcut", Interactive foreground extraction using iterated graph cuts. ACM Trans. Graph., 23(3):309–314, 2004
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top