CODEWITHSUNDEEP
image-processingmachine-learningcomputer-vision
user2867655
user2867655

Reputation: 113

Segmentation in Bag of Word Model

Can someone highlight to me whether the bag of word model performs segmentation process? As far as I know, the BoW treats image features as a set of words just like in documents and represents the image by a histogram or codewords. but I am still confused if the segmentation process is involved? and if it is not? what happen if we do segmentation, I am quite newbie to the field

Upvotes: 1

Views: 875

Answers (1)

G453
G453

Reputation: 1456

I hope these links will help you... Yes, you can perform segmentation with bag of features or words, here is a paper. Link1

This paper “Combined Object Categorization and Segmentation with an Implicit Shape Model” is one of the most famous paper. Even though it is not directly related to your question but go through the related work section of this paper, where they presented many algorithms which uses bag of words sort of techniques for segmentation. It is one of the finest CVPR papers I ever come across. Link2

But bag of words or features are mostly popular in Object Recognition filed only. It is up to you to decide Whether to use it directly or combined it with other algorithms. But i am sure that visual vocabulary formed from local texture descriptors can definitely segment the object of interest.

Upvotes: 4

Related Questions