For example, how can label matrix/vector be built for svm using the features vector got from histogram of segmented image?
your matrix is [m x n], where m is the number of objects (parts in segmented image), and m is a number of features from histogram ?
For each image segment you should have vector with histogram features. This will be a object. Object must have label [for example integers numbers]. So your matrix (learning set) should be [m x n], where m - numbers of objects, n - lenght of features vector plus class label
Thank you. How can I have vector with histogram features for each image segment?
First, can you show a example of segmented results? Can you find each segmented part? (for example by CCL algorithm ?)
How can one come to know which feature from feature vector extracted from histogram, belongs to which segmented category on the basis of which one can construct label matrix for svm?
- Badges
- Science topic
- Similar topics
- Filing