SEGMENTATION

Question

dhanalakshmi 2012-1-5

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/25367-segmentation

编辑： Image Analyst 2023-2-20

hi am doing project in automatic volume measurement of urinary bladder using ultrasound images using image processing . am trying out segmentation but i have a problem that my background and my region of interest are in same color. pls give me some idea n sample code get my roi alone automatically. sample image is in link <http://www.google.co.in/imgres?q=urinary+bladder+image+of+ultrasound+image&um=1&hl=en&sa=N&tbm=isch&tbnid=mVay97b56Hp0IM:&imgrefurl=http://www.medison.ru/uzi/eho400.htm&docid=8Tj1qHtWTbAzdM&imgurl=http://www.medison.ru/uzi/img/p400.jpg&w=640&h=480&ei=bnwFT4KuFIyiiAe3pMS8AQ&zoom=1&iact=hc&vpx=188&vpy=155&dur=2549&hovh=194&hovw=259&tx=182&ty=93&sig=102983391304477840786&page=1&tbnh=146&tbnw=195&start=0&ndsp=20&ved=1t:429,r:0,s:0&biw=1280&bih=832> my roi is center black region.pls any one help me..

3 个评论
显示 1更早的评论隐藏 1更早的评论

dhanalakshmi 2012-1-7

DINESH AND MYSELF ARE WORKING ON SAME PROJECT.

Image Analyst 2012-1-7

I expect that by now you've implemented successfully the simple algorithm I listed below, so go ahead and mark the question as solved.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2012-1-5

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/25367-segmentation#answer_33280

编辑：Image Analyst 2023-2-20

Try this:

Threshold for bright stuff
call imfill
retain the largest blob - that will be your mask.
multiply your mask by the image to get rid of the annotation.
threshold for the dark stuff - it will be arbitrary where you want to define the edges of the bladder.
call bwlabel
call regionprops
sort the measurements by area
the measurements of the largest blob will be your bladder measurements.

See my Image Segmentation Tutorial in my File Exchange:

Image Analyst's File Exchange

It's a generic, general purpose demo of how to threshold an image (of coins) to find blobs, and then measure things about the blobs, and extract a subset blobs based on their areas or diameters.

2 个评论
显示无隐藏无

dhanalakshmi 2012-1-9

thank you.it works

daniel 2013-10-26

hi, could you please publish your code for ultrasound image segmentation?

i would like to extract specific part from the brain cortex of us image, and i wonder if it work for me.

请先登录，再进行评论。

Answer 2

slama najla 2012-5-24

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/25367-segmentation#answer_48907

hello Can you help me? can you give me a exemple code in matlab for region of interest because i need this code very urgent thank

4 个评论
显示 2更早的评论隐藏 2更早的评论

slama najla 2012-5-24

移动：DGM 2023-2-20

thank but i like to give me a code matlab for region of interest thank

Image Analyst 2012-5-25

编辑：Image Analyst 2023-2-20

I did. Did you overlook my link on the segmentation tutorial? It picks out the coins. Each coin can be considered an ROI.

See my Image Segmentation Tutorial in my File Exchange:

Image Analyst's File Exchange

It's a generic, general purpose demo of how to threshold an image to find blobs, and then measure things about the blobs, and extract certain blobs based on their areas or diameters.

请先登录，再进行评论。

Answer 3

Matt Kindig 2012-5-24

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/25367-segmentation#answer_48926

There is no generic code that we can send you--it entirely depends upon your image. Regions of interest can be identified using a number of different methods. Generally the approach is something like this:

- threshold the image to get logical (black&white) image

- perform some cleanup operations (dilation, filtering, etc.)

- use regionprops to identify distinct regions

- identify the desired region.

If you have the Image Processing Toolbox, the demos and examples in the Toolbox help show how to do all of these tasks. These should get you started.

P.S. By the way, you should start a new question rather than simply adding to this thread. Since this thread already has an accepted answer, a lot of people will simply skip over your question (overlooking the time stamp). Once you do that, however, continue to use that same question until your question is resolved, and then accept any answer.