How to extract highest intensity area from a greyscale spectrogram?

4 vues (au cours des 30 derniers jours)
Claudio Eutizi
Claudio Eutizi le 22 Jan 2021
Commenté : Claudio Eutizi le 23 Jan 2021
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.

Réponse acceptée

Image Analyst
Image Analyst le 23 Jan 2021
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 commentaires
Image Analyst
Image Analyst le 23 Jan 2021
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.
Claudio Eutizi
Claudio Eutizi le 23 Jan 2021
Thank you so much.

Connectez-vous pour commenter.

Plus de réponses (1)

KALYAN ACHARJYA
KALYAN ACHARJYA le 22 Jan 2021
Modifié(e) : KALYAN ACHARJYA le 22 Jan 2021
  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 commentaires
KALYAN ACHARJYA
KALYAN ACHARJYA le 23 Jan 2021
Modifié(e) : KALYAN ACHARJYA le 23 Jan 2021
"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.
Claudio Eutizi
Claudio Eutizi le 23 Jan 2021
Yes I found that pixel you're talking about.
I managed to do what I asked to in this question.
Can you help me with this? Thank you so much.

Connectez-vous pour commenter.

Catégories

En savoir plus sur Image Processing and Computer Vision dans Help Center et File Exchange

Produits


Version

R2020b

Community Treasure Hunt

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

Start Hunting!

Translated by