PS求大牛指点，挤牙膏之作。。。

教授的人名等一些个人信息先删掉了。。。感谢各位大牛及热心人提出批评和修改意见。目前字数728.。。

Statement of Purpose

My passion for engineering originated from my uncle who was a PhD of computer science in Canada. He demonstrated his superb ability of programming and operating hardware when I was fifteen. It was that impressive demonstration that made me believe working in a research group and publishing paper as a student would be the coolest thing in four-year college life. Luckily I met with Pro。at the beginning of my second semester in in 2007. He was interested in my treatment with astronomical photos and gave me an opportunity to work in his lab. There I learnt about computer vision, pattern recognition and image processing in a professional way and started my research experience.

During my three years working as a research assistant, I participated in four projects which were “Intelligent Algorithms for Pavement Distress Detection”, “Application of Multi-scale Geometric Transform in Image Processing”, “Application of Contourlet Transform in Image Processing” and “Tensor Voting Algorithm and Application in Remote Sensing Technology”. At that time Wavelet transform had already been applied successfully in JEPG2000, Multi-scale Geometric Transform (MGT) was regarded as the next generation of Wavelet by some researchers. Taking Classified Hidden Markov Tree model as a carrier, my main responsibilities in these projects included developing image denoising, fusing and compression algorithms based on wavelet transform and MGT and making comparisons among Wavelet, Bandelet, Curvelet, Ridgelet and Contourlet. Contourlet was my focus. Most algorithms achieved a higher PSNR value or better vision effect though using contourlet transform in image compressing was tested to be a bad choice. The tensor voting project belonged to Pro. . It’s my favorite project but I felt sorry for not finishing it as I planned. I designed an adaptive 2-D Tensor Voting algorithm which combined with texture spectrum, but I graduated before I finished a version for 3-D. I also had a research on spectrum clustering based on semi-supervised learning, but not too deep.

My research experience started with some obstacles because of lacking enough background accumulations in both theoretical and practical parts. I spent a lot of time to study, to program and to solve every problem I encountered in library and sacrificed most of my weekends and holidays to make up my loss. As a result, an article I wrote was accepted by Laser Journal (EI), which realized my dream of having one paper published as an undergraduate. It further consolidated my desire to become a PhD (like my uncle) in the future. Shortly thereafter, another article “A Static Image Coding Algorithm Based on Contourlet Classified Hidden Markov Tree Model” was published in a British journal, Engineering Computations (SCI).

Considering my experience, computer vision, pattern recognition and image processing are my ideal areas of interest. From my supervisors I know that school of graduate prefers students with strong ability of using professional tools skillfully, solving problems independently and cooperating with other members harmoniously. I prepared myself through years of education and practice in university. I have taken plenty of courses with passion and endeavor and I was awarded fellowship for excellent grade in final exams every year; Out of my curiosity and enthusiasm, I familiarize myself with several programming languages and tools such as C++, Matlab and FPGA(VerilogHDL and VHDL), not to mention the usage of some popular toolboxes and open source libraries. I founded Image Processing Group in faculty with two other students. The group was popular among junior and senior students soon.

As to the graduate school in the future, on one side I plan to study the professional courses systematically, on the other side, I would like to participate in some practical projects with valued applications. The purpose of my overseas study is not only far limited to earn my degree, but also to cultivate myself to be an expert and make contribution to the field and society. I wish to be called Professor or Dr.  in the future. I’ve ever had an idea from the plots in the movie “2012”, which enables the camera to capture the location and track the person we want in a three-dimensional space by distinguishing faces. This system involves several sub-systems, such as database, feedback and face distinguishing module. If constructed, it would play an important role in society. I wish I would have the ability to contribute to such a project in the future.

额，没想到放上来字变得这么小。。。

从小到大怕极了写作。。。但是看了中介写的觉得她小时候一定比我还怕写作。。。。任何建议都可以啊

结构再清晰点吧。。。

比如弄个小标题。。你不是有这么多project。正好合适。

然后多用数字来说话，从第一段开始。

一般我都建议第一段直接提提你对什么方向感兴趣。

最严重的问题，也是个很普遍的问题，中国人的习惯是一堆介绍，然后写核心结论，也就是中心句。比如你第一个project的中心句。I designed an adaptive 2-D Tensor Voting algorithm which combined with texture spectrum。这是第一句就该说的。最精彩的要突出嘛。

写了那么多argument和issue，应该知道，米国人一般第一句是中心句。

我觉得论文名字这些都放在CV里面了，可以不说名字了；多说你的研究经历，成果就适当描述吧；然后是没有结合你申请的学校和department诶，我也在写SOP，国外一个学长说这方面要多结合你申请的department和教授谈谈你未来规划。

nooneknow 发表于 2012-11-22 07:22
结构再清晰点吧。。。

比如弄个小标题。。你不是有这么多project。正好合适。

谢谢，立马改

mamacjj 发表于 2012-11-22 09:55
我觉得论文名字这些都放在CV里面了，可以不说名字了；多说你的研究经历，成果就适当描述吧；然后是没有结合 ...

对的，昨晚也发现了这个问题，就在后面补充了和学校相关的东西，申请的时候会针对不同学校作修改的，谢谢啦