模式识别项目中需要的指针
大家好，我正在做最后一年的项目，我的目标是拥有一个能够识别简单图案的系统，例如条纹、圆点、格子和格子（有点难）。问题是我真的不知道从这里开始。我不确定它属于哪个主题领域（我分为神经网络和人工智能，我将在做该项目时学习这两者）。关于主题领域、编程语言等方面的任何想法都应该从哪里开始。
PS 提前感谢您提供的任何帮助/指示，非常感谢。
Hey guys, am doing a final year project, and my aim it to have a system thats able to recognize simple patters, example stripes, polka dots, checked, and plaid (a bit harder). problem is i dont really know here to start. Am not sure which subject area it falls in (am split between neural nets and AI, both of which i will be learn while doing the project). Any ideas as to where is should start, in terms of subject area, programming language etc.
P.S. Thanks in advance for any help/pointers it's much appreciated.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
你的项目意味着你可能会进行一些人工智能类型的学习。 LISP 和 Prolog 传统上用于此类应用程序。
我不知道你是否必须使用神经网络，但如果你这样做，那么理想情况下你想看看是否能为它们找到一个好的库，然后选择最适合你的技能并允许你使用该库的语言= ）
关于主题，你的教授。可能是询问起点的最佳人选。
PS：神经网络手动实现也不太难，我在我的项目中使用了一个简单的 C++ 语言。
祝你好运！
Your project implies that you will probably be doing some AI type learning. LISP and Prolog have traditionally been used in such applications.
I don't know if you have to use neural nets but if you do, then ideally you want to see if you can find a good library for them and then pick the language that best meets your skills and allows you to use the library =)
Regarding the subject matter, your prof. is probably the best person to ask for starting points.
PS: Neural nets are not too hard to implement manually either, I used a simple one for my project which was in C++.
Good luck!
为了区分条纹、圆点等，我建议基于简单的图像统计数据构建一个分类器。分类器可以是很多东西中的任何一个（神经网络、线性判别式、二次判别式等等），但我希望它们中的任何一个都能够很好地工作。因此，我建议寻找工具来用您已知的语言或工具构建分类器。
To discern among stripes, polka dots, etc., I would suggest building a classifier based on simple image statistics. The classifier could be any of a large number of things (neural network, linear discriminant, quadratic discriminant, blah blah), but I'd expect any of them to work reasonably well. Hence, I suggest looking for tools to build classifiers in the languages or tools you already know.