Regular Articles

Saliency field map construction for region-of-interest-based color image querying

[+] Author Affiliations
Mehmet Celenk

Ohio University, School of Electrical Engineering and Computer Science, Stocker Center, Athens, Ohio 45701

Qiang Zhou

Ohio University, School of Electrical Engineering and Computer Science, Stocker Center, Athens, Ohio 45701

Vermund Vetnes

Ohio University, School of Electrical Engineering and Computer Science, Stocker Center, Athens, Ohio 45701

Rakesh Godavari

Ohio University, School of Electrical Engineering and Computer Science, Stocker Center, Athens, Ohio 45701

J. Electron. Imaging. 14(3), 033012 (July 19, 2005). doi:10.1117/1.1993626
History: Received December 01, 2003; Revised March 07, 2005; Accepted March 08, 2005; Published July 19, 2005; Online July 19, 2005
Text Size: A A A

Spectral (color) and spatial (shape) features available in pictures are two significant sources of information for content-based retrieval of image databases. The developed adaptive shape transform approach originated from the premise that a two-dimensional (2-D) shape can be recovered from a set of Radon-transform-based projections. For search consistency, it is necessary to identify the region(s) of interest (ROI) before applying the Radon transform to the shape query. ROIs are detected automatically by means of saliency-map-based segmentation. The Radon transform packs the shape information of a 2-D object along the projection axes of known orientation, and generates a series of one-dimensional (1-D) functions from color channels for projection angles ranging from 0 to 180 deg. The optimal number of projections for a particular shape is determined by imposing the Kullback-Leibler divergence (KLD) distance measure as the similarity metric between the query and database images. The Radon transforms with the shortest and longest lengths yield the most distinctive shape attributes for the object classes being queried and enable the feature space to be invariant to translation and rotation in the spatial plane. The proposed algorithm is tested on a wide range of color images with complex shaped objects and different spatial resolutions. The KLDs between two images are calculated in the longest and shortest directions of the Radon transform, and then are summed together to find the similarity measure between the query and database pictures. Experimental results are close to those that a human observer expects. Further, the method is quite robust and it can account for high image noise (Shao and Celenk, 2001).

Figures in this Article
© 2005 SPIE and IS&T

Citation

Mehmet Celenk ; Qiang Zhou ; Vermund Vetnes and Rakesh Godavari
"Saliency field map construction for region-of-interest-based color image querying", J. Electron. Imaging. 14(3), 033012 (July 19, 2005). ; http://dx.doi.org/10.1117/1.1993626


Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Related Book Chapters

Topic Collections

Advertisement
  • Don't have an account?
  • Subscribe to the SPIE Digital Library
  • Create a FREE account to sign up for Digital Library content alerts and gain access to institutional subscriptions remotely.
Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).
Access This Proceeding
Sign in or Create a personal account to Buy this article ($15 for members, $18 for non-members).
Access This Chapter

Access to SPIE eBooks is limited to subscribing institutions and is not available as part of a personal subscription. Print or electronic versions of individual SPIE books may be purchased via SPIE.org.