二維繪畫之深度雕刻

Depth Sculpturing of 2D Painting

蔡明宏(Ming-Hung Tsai)

指導教授 ： 吳家麟

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2012年)

https://doi.org/10.6342/NTU.2012.03135

摘要

近幾年來由於3D技術的成熟，真實場景的深度資訊可以直接由深度相機取得，但是已經存在的繪畫若是想要為其附加深度值，唯一的方式只能由使用者來標記。我們提出了一個低時間複雜度的互動式系統來解決這個問題。以往的相關研究大多是採用整張圖一次處裡，minimize energy function的形式來進行，因此十分的耗費時間，而我們的系統將這個問題轉換成filter base algorithm以達到低response time與合理滿意的結果。我們的靈感來自於雕刻的行為，以重複的刻畫行為來進行，在刻畫的過程當中使用者也可以即時的檢視半成品的狀態。由於我們的系統能夠做到即時產生深度圖，藉由投影2D加上深度的資訊到3D當中，可以做出類似版畫的觀察模式來更直覺地掌握現在的情況，如此一來使用者就可以用這個資訊進行判斷下一步的操作。最後我們將展示出擁有depth map之後除了能產生3D視覺效果之外，還能夠藉由這些深度的資訊讓圖片編輯、區域強調甚至轉變成動畫這些原本複雜的行為變得更簡單。

關鍵字

深度圖 ； 雙向濾波器 ； 分水嶺切割法

並列摘要

In real word, the depth of objects can be directly acquired by existing depth cameras, however, the depth information of 2D paintings can only be generated by users. In order to deal with the depth generation problem, a novel and low complexity interactive depth generation approach for 2D paintings is devised. In contrast to traditional approaches which usually addressed this problem through a time-consuming global optimization framework, we formulate the problem as filter-based schemes to achieve reasonable interactive response time. Inspired by sculpturing, the depth information generation can also be addressed as an iterative stroking-and-viewing process. Our work achieves instant response for interactive generating of depth information in which immediate visualization of the depth generation results is possible. Users can then rapidly rectify the current depth generation results. Finally, we illustrate that the newly added depth information for 2D paintings can not only be applied to view 3D effects but also support interesting applications like editing, enhancement and user-controlled animation of 2D paintings.

並列關鍵字

Depth Map ； Bilateral Filter ； Watershed

參考文獻

[2] M. Bertalmio, G. Sapiro, V. Caselles, and C. Ballester. Image inpainting. In Proceedings of the 27th annual conference on Computer graphics and interactive

[3] Y. Boykov and M.-P. Jolly. Interactive graph cuts for optimal boundary amp; region segmentation of objects in n-d images. In Computer Vision, 2001. ICCV 2001. Proceedings. Eighth IEEE International Conference on, volume 1, pages 105 –112 vol.1, 2001.

[6] A. L. Dani, D. Lischinski, and Y. Weiss. Colorization using optimization. ACM Transactions on Graphics, 23:689–694, 2004.

[7] P. F. Felzenszwalb and D. P. Huttenlocher. Efficient graph-based image segmentation. Int. J. Comput. Vision, 59(2):167–181, Sept. 2004.

[9] K. He, J. Sun, and X. Tang. Single image haze removal using dark channel prior. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 33(12):2341 –2353, dec. 2011.