图像风格化技术研究及平台开发
[Abstract]:Animation industry has become a new industry represented by cartoon, online games, mobile phone games and multimedia products, and is regarded as the most promising sunrise industry in the creative economy of the 21st century. The development of animation industry, must be focused on, focus on development can better concentrate and use resources. At present, there is still a lack of animation software in the face of family leisure and entertainment. This platform is mainly aimed at ordinary users to stylize the video. Existing commercial software, such as Photoshop,Premiere Pro, is mainly for professional art workers, through complex editing operations to form a variety of artistic effects. The main research result of this platform is to establish a one-button cartoon special effect model. Through this model, the system can output the artistic effect with a specific style by simply inputting the real image or video and specifying the corresponding special effect type. The main work of this paper is as follows: (1) A fast image sketch generation algorithm based on gray scale synthesis is proposed to realize the artistic sketch generation of realistic images. Different from the existing algorithms, the classical edge detection method is used to extract the edge features of the image, the detailed features of the sketch image are controlled by the distribution of gray information, and the edge features of the image are enhanced and softened by gray level synthesis. (2) establish the color model of cartoon style, describe how to analyze the color space of the original image, remove the color detail feature, reduce the color component, and realize the color abstraction. Using nonlinear filtering algorithm, the filtering technique is essentially a color space clustering model. A typical advantage of nonlinear filtering is that it can effectively preserve contour features. Through nonlinear filtering, not only the unimportant color information can be effectively eliminated, but also the important features such as the contour of the original image can be preserved. In order to achieve the effect of color card. On the basis of the above algorithm and model research, this paper uses VC development tools, combined with OpenCV image processing package, to develop the image style special effect prototype system, and to verify the effectiveness of the above algorithm.
【学位授予单位】:浙江工业大学
【学位级别】:硕士
【学位授予年份】:2010
【分类号】:TP391.41
【参考文献】
相关期刊论文 前10条
1 欧阳鑫玉,赵楠楠,宋蕾,谢元旦;图像分割技术的发展[J];鞍山钢铁学院学报;2002年05期
2 ;从电影到动画 Video Tooning:神奇的风格转换器[J];新电脑;2005年09期
3 梅跃松;杨树兴;莫波;;基于Canny算子的改进的图像边缘检测方法[J];激光与红外;2006年06期
4 王耀贵;;图像高斯平滑滤波分析[J];计算机与信息技术;2008年08期
5 方建文;于金辉;姚t$;彭韧;;利用GPU对3D角色模型绘制水墨效果[J];计算机辅助设计与图形学学报;2007年04期
6 陈博;;我国动漫产业发展中存在的问题与对策[J];齐鲁艺苑;2007年03期
7 陈洪,郑南宁,梁林,李岩,徐迎庆,沈向洋;基于图像的个性化卡通系统[J];软件学报;2002年09期
8 祝长生;王志良;宋倩霞;邓蓉蓉;;表情识别的图像预处理和特征提取方法研究[J];小型微型计算机系统;2009年06期
9 陆宗骐,梁诚;用Sobel算子细化边缘[J];中国图象图形学报;2000年06期
10 冯捷;郑河荣;;基于灰度合成的图像素描效果生成算法[J];浙江工业大学学报;2009年03期
相关博士学位论文 前1条
1 朱胜利;Mean Shift及相关算法在视频跟踪中的研究[D];浙江大学;2006年
相关硕士学位论文 前10条
1 陈方贵;计算机原理及接口实验系统[D];武汉理工大学;2005年
2 吴灵均;非真实感图像绘制中若干算法研究[D];南京理工大学;2005年
3 蒋凌琳;计算机辅助生成剪纸形象[D];广西师范大学;2005年
4 刘丽娜;计算机剪纸纹样自动定位算法研究[D];广西师范大学;2006年
5 李远刚;基于色彩模型的人脸检测方法研究[D];国防科学技术大学;2007年
6 桂斌;基于笔刷的非真实感绘制的研究[D];南京理工大学;2008年
7 姜旭;视频图像处理技术在智能交通系统中的应用[D];苏州大学;2009年
8 刘佳;三维激光全景扫描系统误差分析及相机标定[D];首都师范大学;2009年
9 左丹;基于特征的图像/视频铅笔画绘制技术研究[D];浙江工业大学;2009年
10 胡文锦;图像边缘检测方法研究[D];北京交通大学;2009年
,本文编号:2267661
本文链接:https://www.wllwen.com/wenyilunwen/dongmansheji/2267661.html