利用雲端運算建構基於粒子群最佳化演算法之圖文合成系統

Applying Cloud Computing and Particle Swarm Optimization for Automatically Composing Text-overlaid Images

陳昱叡(Yu-Ruey Chen)

指導教授 ： 洪政欣

國立暨南國際大學/科技學院/資訊工程學系/碩士(2010年)

https://doi.org/10.6837/NCNU.2010.00291

摘要

在先前Lai [12]基於影像美感計算之智慧型圖文合成系統的研究發現，無論是水平平衡、水平對稱、垂直平衡、垂直對稱以及放射性對稱，所設計出來的計算模型都符合人的感知，更發現水平平衡與垂直平衡的平均和人的美感是成正相關性的，套用美感計算模型，一個基於「粒子群最佳化演算法」來計算最佳文字擺放位置的最佳化引擎已經被開發出，並且建立出一套智慧型的圖文合成系統。
本論文主要延伸Lai [12]的研究，並找尋更多的美感向度以及對圖片越大而導致計算時間變長的問題提出探討，首先美感向度中，針對圖文影像中密度的特性，提出密度的計算模型，藉由實驗發現由我們所設計的密度公式中，文字面積占背景區塊面積的密度值約0.12最為符合人的美感，證實密度與美感之間是確實存在著關係；論文並針對圖片越大計像量越大而導致計算時間變長的問題提出解決方式，透過將原先的智慧型的圖文合成系統移置雲端分散式運算平台，藉由平行化處理，並經實驗找出合於雲端的最佳化演算法粒子群參數及MapReduce中的Mapper數，進而加速運算處理時間提高整體系統效率。

關鍵字

美感量化 ； 雲端運算 ； 粒子群演算法 ； 圖文合成系統

並列摘要

Earlier study by Lai et al. [12] developed computation models for visual balances and symmetries for text-overlaid images. The average value of horizontal balance and vertical balance was found to be positively correlated to human’s perception on visual aesthetic. Based on the computation model of visual aesthetics, an optimization engine using "Particle Swarm Optimization (PSO)" was developed to calculate the optimum placement of the text in a text-overlaid image.
This study mainly extends the study by Lai et al. [12]. A computation model for visual density is developed. Experiment results showed that the optimal visual density for text-overlaid images is about 0.12. In addition, this study implemented a Cloud-based the intelligent composition system for text-overlaid images. Evaluation experiment results showed that, by applying the MapReduce parallel computing technique for PSO algorithm, the computation time of the optimization tasks were reduced significantly.

並列關鍵字

Computational aesthetics ； Cloud Computing ； Particle Swarm Optimization ； Automatically Composing Text-overlaid Images

參考文獻

[1] Bauerly, M., Liu, Y., 2006. Computational modeling and experimental investigation of effects of compositional elements on interface and design aesthetics. International Journal of Human-Computer Studies 64(8), 670–682.

[2] Chang, D., Dooley L., Tuovinen, J.E., 2001. Gestalt theory in visual screen design - a new look at an old subject, In Proceedings of the World Conference on Computers in Education 7th, Denmark, 2001.

[3] Chang, F., Dean, J., Ghemawat, S., Hsieh, W., Wallach, D., Burrows, M., Chandra, T., Fikes, A., Gruber, R., 2006. Bigtable: a distributed storage system for structured data. In Proceedings of the 7th USENIX Symposium on Operating Systems Design and Implementation, June. 2006.

[4] Comaniciu, D., Meer, P., 1997. Robust analysis of feature space: color image segmentation. In Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Puerto Rico, June 1997, 750-755.

[5] Dean, J., Ghemawat, S., 2004. MapReduce: simplified data processing on large clusters. In Proceeding of the 6th Symposium on Operating Systems Design and Implementation, December 2004, 137-150.