Abstract—Outdoor scene analysis is a complex problem for
both image processing and pattern recognition domains. There
are two methods of segmenting images to look for objects in an
image, block-based and region-based. Region-based method can
provide some useful information about objects even though
segmentation may not be perfect. There are three phases in this
system: segmentation, features extraction and classification.
The basic idea of this system is to classify local image regions
into semantic concept classes such as tree, sky and road etc. In
this paper, modified Marker-Controlled Watershed (MCWS)
algorithm is proposed. Firstly, the modified (MCWS) algorithm
is used to segment input image. And then, texture feature
vectors are extracted from segmented regions by Gray-Level
Co-occurrence Matrix (GLCM). Finally, classification is
performed by 3-layer Artificial Neural Network (ANN). This
system is applied on real scene images dataset.
Index Terms—Marker-controlled watershed, outdoor scene
analysis, texture.
Kyawt Kyawt Htay1is with University of Computer Studies, Mandalay,
Myanmar (e-mail: kyawtkyawthtay@gmail.com).
Nyein Aye2 is with the head of the Department of Hardware, University
of Computer Studies Mandalay, Myanmar (e-mail:nyeinaye@gmail.com).
[PDF]
Cite: Kyawt Kyawt Htay and Nyein Aye, "Semantic Concepts Classification on Outdoor Scene
Images Based on Region-Based Approach," International Journal of Future Computer and Communication vol. 3, no. 6, pp. 427-431, 2014.
