A new approach to image segmentation with two-dimensional hidden Markov models
Image segmentation is one of the fundamental problems in computer vision. In this work, we present a new segmentation algorithm that is based on the theory of twodimensional hidden Markov models (2D-HMM). Unlike most 2DHMM approaches we do not apply the Viterbi Algorithm, instead we present a computationally efficient algorithm that propagates the state probabilities through the image. This approach can easily be extended to higher dimensions. We compare the proposed method with a 2D-HMM standard algorithm and Iterated Conditional Modes using real world images like a radiography or a satellite image as well as synthetic images. The experimental results show that our approach is highly capable of condensing image segments. This gives our algorithm a significant advantage over the standard algorithm when dealing with noisy images with few classes.
| Main Authors: | , , , |
|---|---|
| Format: | conferenceObject biblioteca |
| Language: | eng |
| Published: |
2013
|
| Subjects: | Classification, Agriculture, Markov Models, Hidden Markov chains, |
| Online Access: | http://hdl.handle.net/11086/21146 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Image segmentation is one of the fundamental
problems in computer vision. In this work, we present a new
segmentation algorithm that is based on the theory of twodimensional hidden Markov models (2D-HMM). Unlike most 2DHMM approaches we do not apply the Viterbi Algorithm, instead
we present a computationally efficient algorithm that propagates
the state probabilities through the image. This approach can
easily be extended to higher dimensions. We compare the proposed method with a 2D-HMM standard algorithm and Iterated
Conditional Modes using real world images like a radiography
or a satellite image as well as synthetic images. The experimental
results show that our approach is highly capable of condensing
image segments. This gives our algorithm a significant advantage
over the standard algorithm when dealing with noisy images with
few classes. |
|---|