臺灣博碩士論文加值系統

人類對於同一種類上的模型,有一種抽象的判斷標準。例如人類對於不同大
小、不同風格的椅子都能成功的認知為椅子,是由於對椅子有一個概念上的認
識。同樣的,人類對於其他物件也會有一個對應上的認知。然而,對於椅子概念
上的認識是什麼?當要具體化的描述時,就面臨了許多挑戰。例如四隻腳的椅
子、滾輪的椅子、單腳的椅子、一整體的椅子、長板凳的椅子等,這些物件人類
總是能知道是椅子,但是判斷的標準是什麼?

基於上述的知覺問題,椅子之所以是椅子是由於人類對於椅子有一個代表性
的認知,通常人類不會把香蕉歸類成一種椅子。然而每個人在對各種物件的代表
性認知與想法卻是不盡相同,因此我們限縮在於日常生活中容易見到的種類,並
進行三維模型的形狀特徵計算,以找出人類對於該物件的代表性判斷觀點。

本研究主要透過三維模型的特徵,如曲率、顯著點等等的幾何特徵;最後透
過機器學習來獲取到人類對於三維模型的判斷與預測,並探討三維模型的形狀特
徵對於人類知覺上的判斷的關聯性。

Human have an abstract cognition on a variety of 3D models. For example, it is easy for us to identify different kinds of chairs as chairs, because we know some concept about the chair. Similarly, for the other type of things we also have some concept to understand them. However, a question simple raises: what is the concept of a chair? It is very challenging when we try to describe it concretely. For instance, we know there are different kinds of chairs, like four-legs chair, swivel chair, stool, rocking chair, single-leg chair and so on. We all know these are chairs, even with different styles or weird structures, but what how do we make our judgment?

According to perception, we discovered that human generally do not treat a banana as a chair, and it is because we know some traits of a chair. But there are different view points from many people, so we just specify some categories that are commonly used in our life. We try to find out the human perspective through the feature computation and learning.

Our study uses different kinds of feature extracted from the 3D models, such as features of curvature, saliency and so on. After that we use machine learning to learn the views from some perceptual study. Finally we can predict and analyze the contribution of features.

中文摘要 ... (iii)
英文摘要 ... (iv)
誌謝 ... (v)
1 緒論 ... (1)
1.1 研究動機 ... (1)
1.2 研究目的 ... (2)
1.3 代表性認知 ... (2)
2 相關文獻 ... (3)
2.1 形狀特徵 ... (3)
2.2 知覺研究 ... (6)
2.3 資料導向方法 ... (9)
3 系統流程 ... (15)
4 調查設計 ... (16)
4.1 代表性選擇 ... (16)
4.2 代表性排序 ... (17)
5 特徵計算 ... (20)
5.1 形狀分佈(Shape Distribution) ... (21)
5.2 形狀直徑函數(Shape Diameter Function) ... (22)
5.3 平均測地線距離(Average Geodesic Distance) ... (23)
5.4 曲率(Curvature) ... (23)
5.5 光場描述子(Light Field Descriptor) ... (25)
5.6 環境光遮蔽(Ambient Occlusion) ... (25)
5.7 顯著性(Saliency) ... (26)
6 代表性預測 ... (28)
6.1 輸入形式 ... (28)
6.2 網路架構 ... (29)
7 實驗結果 ... (31)
7.1 預測結果 ... (31)
7.2 特徵貢獻分析 ... (35)
7.3 實驗環境 ... (36)
8 結論與未來展望 ... (38)
參考文獻 ... (38)

[1] Averkiou, M., Kim, V. G., and Mitra, N. J. Autocorrelation Descriptor for Efficient Co-Alignment of 3D Shape Collections. Computer Graphics Forum 35, 1 (2016), 261–271.
[2] Chen, D.-Y., Tian, X.-P., Shen, Y.-T., and Ouhyoung, M. On Visual Similarity Based 3D Model Retrieval. Computer Graphics Forum (2003).
[3] Chen, X., Golovinskiy, A., and Funkhouser, T. A Benchmark for 3D Mesh Segmentation. ACM Trans. Graph. 28, 3 (July 2009), 73:1–73:12.
[4] Chen, X., Saparov, A., Pang, B., and Funkhouser, T. Schelling Points on 3D Surface Meshes. ACM Trans. Graph. 31, 4 (July 2012), 29:1–29:12.
[5] Chollet, F. Keras. https://github.com/fchollet/keras, 2015.
[6] Cole, F., Golovinskiy, A., Limpaecher, A., Barros, H. S., Finkel-stein, A., Funkhouser, T., and Rusinkiewicz, S. Where Do People Draw Lines? Commun. ACM 55, 1 (Jan. 2012), 107–115.
[7] Gatys, L. A., Ecker, A. S., and Bethge, M. Image Style Transfer Using Convolutional Neural Networks. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (June 2016).
[8] Goldfeather, J., and Interrante, V. A Novel Cubic-order Algorithm for Approximating Principal Direction Vectors. ACM Trans. Graph. 23, 1 (Jan. 2004), 45–63.
[9] Johnson, D. B. Efficient Algorithms for Shortest Paths in Sparse Networks. J. ACM 24, 1 (Jan. 1977), 1–13.
[10] Kalogerakis, E., Chaudhuri, S., Koller, D., and Koltun, V. A Probabilistic Model for Component-based Shape Synthesis. ACM Trans. Graph. 31, 4 (July 2012), 55:1–55:11.
[11] Kalogerakis, E., Hertzmann, A., and Singh, K. Learning 3D Mesh Segmentation and Labeling. ACM Trans. Graph. 29, 4 (July 2010), 102:1–102:12.
[12] Kim, V. G., Li, W., Mitra, N. J., Chaudhuri, S., DiVerdi, S., and Funkhouser, T. Learning Part-based Templates from Large Collections of 3D Shapes. ACM Trans. Graph. 32, 4 (July 2013), 70:1–70:12.
[13] Kingma, D. P., and Ba, J. Adam: A Method for Stochastic Optimization. CoRR abs/1412.6980 (2014).
[14] Kontkanen, J., and Laine, S. Ambient Occlusion Fields. In Proceedings of the 2005 Symposium on Interactive 3D Graphics and Games (New York, NY, USA, 2005), I3D ’05, ACM, pp. 41–48.
[15] Kortgen, M., Novotni, M., and Klein, R. 3D Shape Matching with 3D Shape Contexts. In In The 7th Central European Seminar on Computer Graphics (2003).
[16] Lagae, A., and Dutré, P. A Procedural Object Distribution Function. ACM Trans. Graph. 24, 4 (Oct. 2005), 1442–1461.
[17] Liu, T., Hertzmann, A., Li, W., and Funkhouser, T. Style Compatibility for 3D Furniture Models. ACM Trans. Graph. 34, 4 (July 2015), 85:1–85:9.
[18] Lun, Z., Kalogerakis, E., and Sheffer, A. Elements of Style: Learning Perceptual Shape Style Similarity. ACM Trans. Graph. 34, 4 (July 2015), 84:1–84:14.
[19] Milne, L. Feature Selection Using Neural Networks with Contribution Measures, 1995.
[20] Núñez, R., and Freeman, W. Reclaiming Cognition: The Primacy of Action, Intention and Emotion. Journal of Consciousness Studies, 6, No. 11-12. Imprint Academic, 1999.
[21] Osada, R., Funkhouser, T., Chazelle, B., and Dobkin, D. Shape Distributions. ACM Trans. Graph. 21, 4 (Oct. 2002), 807–832.
[22] Rusu, R. B., Blodow, N., and Beetz, M. Fast Point Feature Histograms (FPFH) for 3D registration. In 2009 IEEE International Conference on Robotics and Automation (May 2009), pp. 3212–3217.
[23] Shapira, L., Shamir, A., and Cohen-Or, D. Consistent Mesh Partitioning and Skeletonisation Using the Shape Diameter Function. Vis. Comput. 24, 4 (Mar. 2008), 249–259.
[24] Shtrom, E., Leifman, G., and Tal, A. Saliency Detection in Large Point Sets. In 2013 IEEE International Conference on Computer Vision (Dec 2013), pp. 3591–3598.
[25] Sidi, O., van Kaick, O., Kleiman, Y., Zhang, H., and Cohen-Or, D. Unsupervised Co-segmentation of a Set of Shapes via Descriptor-space Spectral Clustering. ACM Trans. Graph. 30, 6 (Dec. 2011), 126:1–126:10.
[26] Wang, Y., Asafi, S., van Kaick, O., Zhang, H., Cohen-Or, D., and Chen, B. Active Co-analysis of a Set of Shapes. ACM Trans. Graph. 31, 6 (Nov. 2012), 165:1–165:10.
[27] Xu, K., Kim, V. G., Huang, Q., Mitra, N., and Kalogerakis, E. Data-driven Shape Analysis and Processing. In SIGGRAPH ASIA 2016 Courses (New York, NY, USA, 2016), SA ’16, ACM, pp. 4:1–4:38.
[28] Yumer, M. E., Chaudhuri, S., Hodgins, J. K., and Kara, L. B. Semantic Shape Editing Using Deformation Handles. ACM Trans. Graph. 34, 4 (July 2015), 86:1–86:12.