i am looking for a image segmentation algorithm which can extract features from my desire part. ex: edge detection or background reduction. i have used watershed algorithm but can not get proper result. also i used Grabcut method where i find edge detection but not fully. also from there i can not get my features from the desire part. Now a days i am working with K-means clustering method in which i am trying to improving the algorithm. but K-means works in Machine Learning(ML) and for ML it is not possible to apply in one or two images, for this i need minimum 10,000 data. but at this time i do not need that much of data.is there any algorithm available for image segmentation?
You can use colored flooding algorithm, starting from the point on the interesting object. The principle is as follows, a point can be considered part of the object if the difference in intensity with its neighbors is smaller than a treshold you specify.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images