結合場景偵測的人臉與色彩美化

Face Beautification and Color Enhancement with Scene Mode Detection

黃大源(Da-Yuan Huang)

指導教授 ： 傅楸善

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2010年)

https://doi.org/10.6342/NTU.2010.00721

摘要

自動相片美化諸如自動白平衡、消除雜訊、自動人臉美化等等技術隨著數位相機、印表機等數位影像產品的普及，扮演著提昇相片品質的重要角色。
但是，由於使用者所拍得相片各式各樣，人臉美化、照片美化等演算法很難自動去完成。造成這些演算法往往需要使用者近一步手動調整，而無法達到自動美化目的，造成使用者在修改照片參數的時候需要近一步的調整、學習，增加了使用的困難度。
在此篇論文中，我們開發出一個結合自動人臉美化、自動色彩美化兩種功能的演算法，目標是能夠強化背景色彩、細節的豐富程度，又能夠美化照片中人臉的肌膚，並且在短時間內產出穩定且和諧的美化結果，讓我們的演算法能實際實做在印表機上且被使用者使用。最後，我們會依據相片的顏色內容，找出最符合相片的場景模式，讓印表機能夠根據場景資訊使用對應的量變曲線去列印，讓照片的整體顏色更鮮艷漂亮。

關鍵字

人臉美化 ； 色彩美化 ； 場景分類

並列摘要

Auto photo beautification such as automatic white balance, noise reduction, and automatic human face beautification play important roles of promoting the quality of photograph because of the popularity of digital image production such as digital camera and printer.
However, due to input image variety, algorithms such as automatic human face beautification and photo color enhancement are hard to automate. Parameters of algorithm usually have to be adjusted manually, and users have to learn how to adjust them. That increases the difficulty of using those products.
In this thesis, we develop a pipeline which combines automatic human face beautification and automatic color enhancement. The goal of this thesis is to beautify human faces and enhance the detail and color of the background of photos. The enhancement methods must produce steady and harmonious results quickly, so that our methods can be implemented on printer and used by users.
Finally, we will select the best-match scene mode based on the color distribution of input image. The printer can use the corresponding profile with the information of the scene; it will make the print-out photos more colorful and vivid.

並列關鍵字

face ； beautification ； color ； enhancement ； scene ； detection

參考文獻

[2] Anthropics Technology, “Portrait Professional,” http://www.portraitprofessional.com/, 2010.

[4] C. C. Chang, C. J. Lin, “LIBSVM -- A Library for Support Vector Machines”, http://www.csie.ntu.edu.tw/~cjlin/libsvm/index.html, 2010.

[5] C. W. Chen, D. Y. Huang, and C. S. Fuh, “Automatic Skin Color Beautification”, Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, p. 157, 2009.

[8] D. Y. Huang and C. S. Fuh, “Automatic Face Color Enhancement,” Proceedings of IPPR Conference on Computer Vision, Graphics, and Image Processing, Shitou, Taiwan, D6-8, p. 175, 2009.

to Bilateral Filtering and its Applications,” http://people.csail.mit.edu/sparis/bf_course/, 2010.