照片拼貼是一個能夠快速地表達大量資訊內容的呈現方式，將要呈現的所有照片排列在一個特定範圍內，讓使用者可以一次瀏覽全部的內容。然而，大多數的照片拼貼應用多呈現於傳統的四方形畫布，較無藝術性且無法吸引使用者觀看，且由於在排列時使用由上而下的演算法，最佳解往往無法讓使用者隨心所欲的將想要的圖片擺放在一起。本篇基於語義將圖片分群，並透過照片集語義內容分析群與群之間的關聯度，將此相關性以二維距離的表現方式投影在二維平面空間，計算出每張照片的座標位置，將照片擺放在使用者指定的任意形狀畫布上。本篇提出一個新的方法，利用文字雲的概念快速地建立照片拼貼的初步排版，在圖片尺寸上突顯重要的圖片，讓使用者可以優先專注在此拼貼要呈現的重點上。接著本篇提出一個優化排版的演算法，將畫布空間充分利用，產生緊密的照片拼貼結果，並利用影像調整大小的方法降低演算法過程中造成的照片失真，最終呈現出一個令使用者視覺上感到愉快且有趣的照片拼貼作品。

Photo collage is a fast way to represent a huge amount of information. A photo collage shows all the images in a specific shape boundary, users can look through all contents at once. However, most traditional photo collage applications are in the shape of rectangles, which is less artistic and less appealing to users. And most of them use top-down algorithm when solving the optimal solution to the layout, so users cannot place images at a designated place.

We cluster the images base on their semantic contents and analyze the relevance between clusters according to the semantic contents of the image data set. Then we represent the relations using 2D distance and project them in a 2D space. Therefore we can calculate the 2D coordinate of each image and place the image on the canvas of any shape the user specified.

We introduce a new method which uses the concept of word cloud to create the initial layout of the photo collage fast. The photo collage we create is saliency-aware. The more salient image will be highlighted by a bigger size, so users can focus on them first when looking through the whole collage. We then present an algorithm of layout optimization. Fully utilized the area of the canvas, so we can create very compact photo collages. We use image resizing to alleviate image distortions caused during our layout algorithm. Finally, we present a visual-pleasing and interesting photo collage.

[1] Yunhai Wang, Xiaowei Chu, Chen Bao, Lifeng Zhu, Oliver Deussen, Baoquan Chen and Michael Sedlmair, "EdWordle: Consistency-Preserving Word Cloud Editing," IEEE Transactions on Visualization and Computer Graphics, vol. 24, pp. 647-656, 2018.
[2] Bernhard Reinert, Tobias Ritschel and Hans-Peter Seidel, "Interactive By-example Design of Artistic Packing Layouts," ACM Transactions on Graphics, vol. 32, no. 6, 2013.
[3] Kin Chung Kwan, Lok Tsun Sinn, Chu Han, Tien-Tsin Wong and Chi-Wing Fu, "Pyramid of Arclength Descriptor for Generating Collage of Shapes," ACM Transactions on Graphics, vol. 35, no. 229, 2016.
[4] V. Cheung, "Shape Collage," 2007. [Online]. Available: https://shapecollage.com/.
[5] D. Stromi, "ShapeX," [Online]. Available: http://www.reasyze.com/shapex/.
[6] "Ribbet," [Online]. Available: https://www.ribbet.com/.
[7] J. Fan, "Photo Layout with a Fast Evaluation Method and Genetic Algorithm," 2012 IEEE International Conference on Multimedia and Expo Workshops, pp. 308-313, 2012.
[8] Zhipeng Wu and Kiyoharu Aizawa, "PicWall: Photo collage on-the-fly," 2013 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, 2013.
[9] Shiguang Liu, Xiaobing Wang, Ping Li and Junyong Noh, "Trcollage: Efficient Image Collage Using Tree-Based Layer Reordering," 2017 International Conference on Virtual Reality and Visualization, 2017.
[10] Ming-Te Chi, Shih-Syun Lin, Shiang-Yi Chen, Chao-Hung Lin, and Tong-Yee Lee, "Morphable Word Clouds for Time-Varying Text Data Visualization," IEEE Transactions on Visualization and Computer Graphics, vol. 21, no. 12, pp. 1415 - 1426, 2015.
[11] C. B. Atkins, "Blocked Recursive Image Composition," MM '08 Proceedings of the 16th ACM international conference on Multimedia, pp. 821-824, 2008.
[12] Xingjia Pan, Fan Tang, Weiming Dong, Chongyang Ma, Yiping Meng, Feiyue Huang,Tong-Yee Lee and Changsheng Xu, "Content-Based Visual Summarization for Image Collection," IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS, vol. XX, no. XX, 2019.
[13] Xintong Han, Chongyang Zhang, Weiyao Lin, Mingliang Xu, Bin Sheng, and Tao Mei, "Tree-based Visualization and Optimization for Image Collection," IEEE Transactions on Cybernetics, vol. 46, pp. 1286-1300, 2016.
[14] Joseph Redmon and Ali Farhadi, "YOLOv3: An Incremental Improvement," arXiv, 2018.
[15] I. Jolliffe, "Principal Component Analysis," Springer Verlag, 1986.
[16] Laurens van der Maaten and Geoffrey Hinton, "Visualizing Data using t-SNE," Journal of Machine Learning Research, vol. 9, pp. 2579-2605, 2008.
[17] Shai Avidan and Ariel Shamir, "Seam Carving for Content-Aware Image Resizing," ACM Transactions on Graphics, vol. 26, no. 10, 2007.
[18] Shih-Syun Lin, I-Cheng Yeh, Chao-Hung Lin, and Tong-Yee Lee,, "Patch-Based Image Warping for Content-Aware Retargeting," IEEE TRANSACTIONS ON MULTIMEDIA, vol. 15, no. 2, pp. 359-368, 2013.
[19] Si Liu, Zhen Wei, Yao Sun, Xinyu Ou, Junyu Lin, Bin Liu and Ming-Hsuan Yang, "Composing Semantic Collage for Image Retargeting," IEEE Transactions on Image Processing, vol. 27, pp. 5032-5043, 2018.
[20] "Photovisi," [Online]. Available: https://www.photovisi.com/.
[21] "Photo Collage," Zygomatic, 2003. [Online]. Available: https://www.photocollage.com/.