A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge

Mehryar Emambakhsh, Hossein Ebrahimnezhad, Mohammad Hossein Sedaaghi

Research output: Contribution to conferencePaper

1 Citation (Scopus)

Abstract

Image segmentation is maybe one of the most fundamental topics in image processing. Among numerous methods for segmentation, blind (bottom-up) algorithms, which are based on intrinsic image features, e.g. intensity, colour and texture, have been used extensively. However, there are some situations such as poor image contrast, noise, and also occlusion that result in failure for blind segmentation methods. Therefore, prior knowledge of the object of interest must be involved in the segmentation approaches. For this purpose, in this work, a novel integrated algorithm is proposed, which is a combination of bottom-up (blind) and top-down (including shape prior) segmentation algorithms. In our approach, after a colour space transformation, an energy function based on non-linear diffusion of colour components and directional derivatives, is defined. After that, some distance maps of the object of interest are generated from binary images that contain the training shapes of the object. Finally, the energy function minimization is done by evolving a level set function, which is set up by the distance maps. The results show our region-based segmentation algorithm robustness against noise and occlusion.

Conference

Conference18th Iranian Conference on Electrical Engineering (ICEE)
CityIsfahan, Iran
Period11/05/1013/05/10

Fingerprint

Image segmentation
Textures
Color
Binary images
Set theory
Image processing
Derivatives

Cite this

Emambakhsh, M., Ebrahimnezhad, H., & Sedaaghi, M. H. (2010). A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge. Paper presented at 18th Iranian Conference on Electrical Engineering (ICEE), Isfahan, Iran, . https://doi.org/10.1109/IRANIANCEE.2010.5507063

A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge. / Emambakhsh, Mehryar; Ebrahimnezhad, Hossein; Sedaaghi, Mohammad Hossein.

2010. Paper presented at 18th Iranian Conference on Electrical Engineering (ICEE), Isfahan, Iran, .

Research output: Contribution to conferencePaper

Emambakhsh, M, Ebrahimnezhad, H & Sedaaghi, MH 2010, 'A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge' Paper presented at 18th Iranian Conference on Electrical Engineering (ICEE), Isfahan, Iran, 11/05/10 - 13/05/10, . https://doi.org/10.1109/IRANIANCEE.2010.5507063
Emambakhsh M, Ebrahimnezhad H, Sedaaghi MH. A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge. 2010. Paper presented at 18th Iranian Conference on Electrical Engineering (ICEE), Isfahan, Iran, . https://doi.org/10.1109/IRANIANCEE.2010.5507063
Emambakhsh, Mehryar ; Ebrahimnezhad, Hossein ; Sedaaghi, Mohammad Hossein. / A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge. Paper presented at 18th Iranian Conference on Electrical Engineering (ICEE), Isfahan, Iran, .
@conference{2536c02e85aa46e780195cd7c918acb4,
title = "A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge",
abstract = "Image segmentation is maybe one of the most fundamental topics in image processing. Among numerous methods for segmentation, blind (bottom-up) algorithms, which are based on intrinsic image features, e.g. intensity, colour and texture, have been used extensively. However, there are some situations such as poor image contrast, noise, and also occlusion that result in failure for blind segmentation methods. Therefore, prior knowledge of the object of interest must be involved in the segmentation approaches. For this purpose, in this work, a novel integrated algorithm is proposed, which is a combination of bottom-up (blind) and top-down (including shape prior) segmentation algorithms. In our approach, after a colour space transformation, an energy function based on non-linear diffusion of colour components and directional derivatives, is defined. After that, some distance maps of the object of interest are generated from binary images that contain the training shapes of the object. Finally, the energy function minimization is done by evolving a level set function, which is set up by the distance maps. The results show our region-based segmentation algorithm robustness against noise and occlusion.",
author = "Mehryar Emambakhsh and Hossein Ebrahimnezhad and Sedaaghi, {Mohammad Hossein}",
year = "2010",
doi = "10.1109/IRANIANCEE.2010.5507063",
language = "English",
note = "18th Iranian Conference on Electrical Engineering (ICEE) ; Conference date: 11-05-2010 Through 13-05-2010",

}

TY - CONF

T1 - A hybrid top-down/bottom-up approach for image segmentation incorporating colour and texture with prior shape knowledge

AU - Emambakhsh, Mehryar

AU - Ebrahimnezhad, Hossein

AU - Sedaaghi, Mohammad Hossein

PY - 2010

Y1 - 2010

N2 - Image segmentation is maybe one of the most fundamental topics in image processing. Among numerous methods for segmentation, blind (bottom-up) algorithms, which are based on intrinsic image features, e.g. intensity, colour and texture, have been used extensively. However, there are some situations such as poor image contrast, noise, and also occlusion that result in failure for blind segmentation methods. Therefore, prior knowledge of the object of interest must be involved in the segmentation approaches. For this purpose, in this work, a novel integrated algorithm is proposed, which is a combination of bottom-up (blind) and top-down (including shape prior) segmentation algorithms. In our approach, after a colour space transformation, an energy function based on non-linear diffusion of colour components and directional derivatives, is defined. After that, some distance maps of the object of interest are generated from binary images that contain the training shapes of the object. Finally, the energy function minimization is done by evolving a level set function, which is set up by the distance maps. The results show our region-based segmentation algorithm robustness against noise and occlusion.

AB - Image segmentation is maybe one of the most fundamental topics in image processing. Among numerous methods for segmentation, blind (bottom-up) algorithms, which are based on intrinsic image features, e.g. intensity, colour and texture, have been used extensively. However, there are some situations such as poor image contrast, noise, and also occlusion that result in failure for blind segmentation methods. Therefore, prior knowledge of the object of interest must be involved in the segmentation approaches. For this purpose, in this work, a novel integrated algorithm is proposed, which is a combination of bottom-up (blind) and top-down (including shape prior) segmentation algorithms. In our approach, after a colour space transformation, an energy function based on non-linear diffusion of colour components and directional derivatives, is defined. After that, some distance maps of the object of interest are generated from binary images that contain the training shapes of the object. Finally, the energy function minimization is done by evolving a level set function, which is set up by the distance maps. The results show our region-based segmentation algorithm robustness against noise and occlusion.

UR - http://dx.doi.org/10.1109/IRANIANCEE.2010.5507063

U2 - 10.1109/IRANIANCEE.2010.5507063

DO - 10.1109/IRANIANCEE.2010.5507063

M3 - Paper

ER -