m*t
2 楼
what you mean position? all the vertices coordinates?
m*s
3 楼
well，i think just height, width and center point. That is good enough.
Thank you.
【在 m***t 的大作中提到】
: what you mean position? all the vertices coordinates?
m*t
4 楼
I am not good at graphics, but here is my suggestion:
there are two steps of solving this problem, first identify all the
distinct shapes in the picture, second calculate the center.
second step is easy, since once you have the shape, you just get mean for
height/width of all pixels in the shape.
first step is basically constructing trees in forest. mark the visited
tree to a third color, say red, and keep going until done.
there are two steps of solving this problem, first identify all the
distinct shapes in the picture, second calculate the center.
second step is easy, since once you have the shape, you just get mean for
height/width of all pixels in the shape.
first step is basically constructing trees in forest. mark the visited
tree to a third color, say red, and keep going until done.
w*p
5 楼
First step is to segment the foreground shape from background. Considering
your case (color fixed), the easiest way is region growing.
Second step of computing the center, do you have prior knowledge about what
shapes are in?
【在 m**s 的大作中提到】
: 如果我有一个bmp文件，里面有很多随机图形（黑底色，图形颜色固定），我如何把那
: 些图形的位置，中心给确定住呢？先谢谢大家了。
your case (color fixed), the easiest way is region growing.
Second step of computing the center, do you have prior knowledge about what
shapes are in?
【在 m**s 的大作中提到】
: 如果我有一个bmp文件，里面有很多随机图形（黑底色，图形颜色固定），我如何把那
: 些图形的位置，中心给确定住呢？先谢谢大家了。
m*s
6 楼
thanks guys.
By the way, could you guys suggest some intro books about dealing with
graphics like that?
By the way, could you guys suggest some intro books about dealing with
graphics like that?
m*s
7 楼
what
Thank you whkylp, i think most of them are round, or ellipses, but my
problem is that i don't have any backgroud about graphic programming. So,
could you suggest me some books to read? Thank you very much!
【在 w****p 的大作中提到】
: First step is to segment the foreground shape from background. Considering
: your case (color fixed), the easiest way is region growing.
: Second step of computing the center, do you have prior knowledge about what
: shapes are in?
w*p
8 楼
This book is like a image processing bible to me: Digital Image Processing,
2nd Edition. See the link of this book: http://www.imageprocessingplace.com/
Hope this helps.
【在 m**s 的大作中提到】
:
: what
: Thank you whkylp, i think most of them are round, or ellipses, but my
: problem is that i don't have any backgroud about graphic programming. So,
: could you suggest me some books to read? Thank you very much!
2nd Edition. See the link of this book: http://www.imageprocessingplace.com/
Hope this helps.
【在 m**s 的大作中提到】
:
: what
: Thank you whkylp, i think most of them are round, or ellipses, but my
: problem is that i don't have any backgroud about graphic programming. So,
: could you suggest me some books to read? Thank you very much!
o*r
9 楼
厚脸皮吱一声，
其实我是搞image processing地，
以前。
而且我还做过你的这个问题。
就如前楼所说，
你要先作segmentation,
方法一把一把的，
那个digital image processing是bible,
很多image segmentation paper也可以参考
然后捏是ellipses recognition
就是拿ellipses方程往你的segmentation结果上套
本来我还想攒个conference paper,
没啥新意才作罢。
【在 m**s 的大作中提到】
:
: what
: Thank you whkylp, i think most of them are round, or ellipses, but my
: problem is that i don't have any backgroud about graphic programming. So,
: could you suggest me some books to read? Thank you very much!
其实我是搞image processing地，
以前。
而且我还做过你的这个问题。
就如前楼所说，
你要先作segmentation,
方法一把一把的，
那个digital image processing是bible,
很多image segmentation paper也可以参考
然后捏是ellipses recognition
就是拿ellipses方程往你的segmentation结果上套
本来我还想攒个conference paper,
没啥新意才作罢。
【在 m**s 的大作中提到】
:
: what
: Thank you whkylp, i think most of them are round, or ellipses, but my
: problem is that i don't have any backgroud about graphic programming. So,
: could you suggest me some books to read? Thank you very much!
m*s
10 楼
,
O, thank you very much!
【在 w****p 的大作中提到】
: This book is like a image processing bible to me: Digital Image Processing,
: 2nd Edition. See the link of this book: http://www.imageprocessingplace.com/
: Hope this helps.
m*s
11 楼
非常感谢，mitbbs果然大牛多！
【在 o******r 的大作中提到】
: 厚脸皮吱一声，
: 其实我是搞image processing地，
: 以前。
: 而且我还做过你的这个问题。
: 就如前楼所说，
: 你要先作segmentation,
: 方法一把一把的，
: 那个digital image processing是bible,
: 很多image segmentation paper也可以参考
: 然后捏是ellipses recognition
相关阅读
Google的Quoc Le有多牛？Re: 码工就是吃青春饭的，我以前的公司有两三天不睡连续干的 (转载)[bssd]万物皆数Java 在ML和DL还不错了gdb vs print or println微软AI 能编程。。。发现了个微软3位同胞的强文 Cat Head Detection来，分析下这个垃圾邮件里面的代码Tensorflow's Mandelbrot Set Tutorial请问Java和Java EE的区别是啥？一个白人老美程序员这么评价老中和老印说个两周前发生的面试故事 (转载)lua怎么样[bssd]Wdong你之前是不是提过王培教授？把web service+client 包装成一个桌面GUI 软件容易吗？谷歌照片翻译那种从图片里识别文字用什么库能做到码农的明天？知乎网页的防复制功能是怎么实现的？能否请推荐个社区系统可以双向同步Slack的?Visual Studio 2013新文件模板