How to extract highest intensity area from a greyscale spectrogram?

6 次查看（过去 30 天）
Claudio Eutizi
Claudio Eutizi 2021-1-22
评论： Claudio Eutizi 2021-1-23
Hello.
I got greyscale mel-spectrograms images from a dataset and I want to divide it into several areas and to obtain the area where pixels have highest intensity in average.
This will be useful to label the images with rectangles for a deep learning training.
Hope somebody will help me.
I attach a greyscale spectrogram I got where you can show me the way to do it.
Thank you.

请先登录，再回答此问题。

采纳的回答

Image Analyst
Image Analyst 2021-1-23
You've given no criteria for how those areas are to be determined. You might want to use watershed() or superpixels(). Or use imbinarize() to segment on intensity (adaptive or global), or multithresh() for several global thresholds.
  3 个评论
显示 1更早的评论
Image Analyst
Image Analyst 2021-1-23
Oh, OK. I would have used blockproc() if you wanted rectangular blocks but glad you solved it somehow.
I'm attaching several blockproc() demos in case you're still interested.

请先登录，再进行评论。

更多回答（1 个）

KALYAN ACHARJYA
KALYAN ACHARJYA 2021-1-22
编辑：KALYAN ACHARJYA 2021-1-22
  1. Apply thresholding to cluster the image into two segment, certain higher pixel and lower value pixels.
  2. Get the largest blob as per requirement. (bwareafilt function)
What does "Major Intensity" mean here?
  5 个评论
显示 3更早的评论
KALYAN ACHARJYA
KALYAN ACHARJYA 2021-1-23
编辑：KALYAN ACHARJYA 2021-1-23
"but this code you wrote here shows a black image".
Most probably, there is only one maximum value pixel, so it is not easily visualized (check carefully). You can confirm the same with the extract_roi matrix, which must be the non-zero matrix.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Image Processing and Computer Vision 的更多信息

标签

产品


版本

R2020b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by