Different criteria can be used in clustering approaches, some include:
Minimize cluster distances
Maximize cluster distances
Cluster by density
Cluster by data type
Answer
Maximize cluster distances
Cluster by density
Card 22
Question
Overlap in defining cluster is used to …
Model uncertainty stemming from imaging (e.g., PVE, poor SNR)
Not really used as clusters are typically well separated
Cope with intensity inhomogeneities
Answer
Model uncertainty stemming from imaging (e.g., PVE, poor SNR)
Card 23
Question
The idea behind fuzzy c-means is to …
Iterate between cluster center and membership value updates
Maximize voxel like hood based on intensity information
Cluster noisy (fuzzy) intensity information using average estimations
Answer
Iterate between cluster center and membership value updates
Card 24
Question
GMM uses the EM algorithm in order to …
To speed up the segmentation
GMM does not use EM
To estimate model parameters and voxel labels
To estimate one set of parameters and voxel labels
Closed form to estimate voxel lables
Answer
To estimate model parameters and voxel labels
How to use this set
Read the preview and check whether the content and answers suit your learning goal. You can add the public set to your sets to study it. Your account shows the available actions.