I noticed that in all the research papers that dealt with Markov Random Fields Fuzzy C-Means (MRFFCM) clustering of synthetic aperture RADAR (SAR) images, the conditional probability is computed based on the normal probability density.
Can anyone explain the reason for this Gaussian function use? Is it related to the data distribution?
Thank you!
Dear Chayma Chaabani,
It seems to me that in order to ensure the homogeneity condition of the conditional probability for all the pixels, the normal distribution function is chosen. Please see paper "Suitable sampling..."
Here is attached files in topics.
Best regards
Thank you for the answer.
According to the files that you have kindly attached, I can see that in the context of MRFFCM clustering, the conditional probability can be considered as normal locally (at the level of the pixel neighborhood) regardless of the probability density function of the whole data.
Is that correct?
- Badges
- Science topic
- Similar topics
- Filing