Periodic Pattern Formation and Its Applications to Image Processing in Cellular Neural Networks
讲座名称：
Periodic Pattern Formation and Its Applications to Image Processing in Cellular Neural Networks
讲座时间：
2011-11-15
讲座人：
Yoshifumi Nishio
形式：
校区：
兴庆校区
实践学分：
讲座内容：
应电信学院自动化系杨国安副教授的邀请，日本德岛大学先进技术研究院Prof.Yoshifumi Nishio来我校做学术报告。
报告题目：Periodic Pattern Formation and Its Applications to
Image Processing in Cellular Neural Networks
报告时间：2011年11月15日(周二)上午9：00
报告地点：西一楼第一会议室
报告摘要：Although the importance of computer vision has been recognized widely, it is still difficult to develop a computer vision system with the performance of animals’ eye. In computer vision, first, a general feature of images is extracted, and next, detailed analyses and image recognition are performed on its basis. However, it is difficult to develop a universal feature extraction method for various kinds of images in real world. Hence, it is important to devise more general feature extraction method, which does not deeply depend on a type of images. We have proposed an interesting method for image processing using periodic pattern formation in the Cellular neural network (CNN). In this talk, First, I introduce two different CNN templates producing periodic patterns. Next, I explain two image processing algorithms using one of the templates, one is the algorithm extracting spatial uniformity and the other is the algorithm segmenting texture by enhancing spatial periodicity. Some simulation results are shown to demonstrate our idea.
相关视频