Image Annotation With Semi-Supervised Clustering

2009-09-16
Methods developed for image annotation usually make use of region clustering algorithms. Visual codebooks are generated from the region clusters of low level features. These codebooks are then, matched with the words of the text document related to the image, in various ways. In this paper, we supervise the clustering process by using three types of side information. The first one is the topic probability information obtained from the text document associated with the image. The second is the orientation and the third one is the color information around each interest point. The side information provides a set of constraints in a semi-supervised k-means region clustering algorithm. Consequently, in clustering of regions not only low level features, but also this extra information is used. Experimental results show that image annotation with semi-supervision of side information is more successful compared to the one that uses low level features alone. Moreover, a speedup is obtained in the modified k-means algorithm because of the constraints. The proposed algorithm is implemented in a high performance parallel computation environment.

Suggestions

Image Annotation by Semi-Supervised Clustering Constrained by SIFT Orientation Information
Sayar, Ahmet; Yarman-Vural, Fatos T. (2008-10-29)
Methods developed for image annotation usually make use of region clustering algorithms. Visual codebooks are generated from the region clusters of low level features. These codebooks are then, matched with the words of the text document related to the image, in various ways. In this paper, we supervise the clustering process by using the orientation information assigned to each interest point of Scale-invariant feature transform (SIFT) features to generate a visual codebook. The orientation information pro...
Image segmentation with unified region and boundary characteristics within recursive shortest spanning tree
Esen, E.; Alp, Y. K. (2007-06-13)
The lack of boundary information in region based image segmentation algorithms resulted in many hybrid methods that integrate the complementary information sources of region and boundary, in order to increase the segmentation performance. In compliance with this trend, we propose a novel method to unify the region and boundary characteristics within the canonical Recursive Shortest Spanning Tree algorithm. The main idea is to incorporate the boundary information in the distance metric of RSST with minor cha...
Feature extraction from acoustic and hyperspectral data by 2d local discriminant bases search
Kalkan, Habil; Kalkan, Habil; Department of Information Systems (2008)
In this thesis, a feature extraction algorithm based on 2D Local Discriminant Bases (LDB) search is developed for acoustic and hyperspectral data. The developed algorithm extracts the relevant features by both eliminating the irrelevant ones and/or by merging the ones that do not provide extra information on their own. It is implemented on real world data to separate aflatoxin contaminated or high risk hazelnuts from the sound ones by using impact acoustic and hyperspectral data. Impact acoustics data is us...
Shape recognition with generalized beam angle statistics
Tola, OO; Arica, N; Yarman Vural, Fatoş Tunay (2004-04-30)
In this study, we develop a new shape descriptor and matching algorithm in order to find a given template shape in an edge detected image without performing boundary extraction. The shape descriptor based on Generalized Beam Angle Statistics (GBAS) defines the angles between the lines connecting each boundary point with the rest of the points, as random variable. Then, it assigns a feature vector to each point using the moments of beam angles. The proposed matching algorithm performs shape recognition by ma...
Image annotation with semi-supervised clustering
Sayar, Ahmet; Yarman Vural, Fatoş Tunay; Department of Computer Engineering (2009)
Image annotation is defined as generating a set of textual words for a given image, learning from the available training data consisting of visual image content and annotation words. Methods developed for image annotation usually make use of region clustering algorithms to quantize the visual information. Visual codebooks are generated from the region clusters of low level visual features. These codebooks are then, matched with the words of the text document related to the image, in various ways. In this th...
Citation Formats
A. Sayar and F. T. Yarman Vural, “Image Annotation With Semi-Supervised Clustering,” 2009, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/35299.