An ROI Search Method for Still Images Based on Set Descriptions

Authors:
Taek Mu Kwon;Pankaj Agrawal;Donald B. Crouch
Affiliations:
Department of Electrical and Computer Engineering, University of Minnesota, Duluth, Duluth, MN 55812;Department of Computer Science, University of Minnesota, Duluth, Duluth, MN 55812;Department of Computer Science, University of Minnesota, Duluth, Duluth, MN 55812
Venue:
Multidimensional Systems and Signal Processing
Year:
1998

Citing 5
Cited 0

Corner detection and curve representation using cubic B-spline

Computer Vision, Graphics, and Image Processing
Human face profile recognition by computer

Pattern Recognition
Algebraic feature extraction of image for recognition

Pattern Recognition
Identification of human faces based on isodensity maps

Pattern Recognition
Contour shape description based on an arch height function

Pattern Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

Detecting a Region of Interest (ROI) from a digitizedimage is an important step towards highly efficient image communication.Although a significant amount of work has been done for timevarying images, in which the ROI is expressed in terms of thedifferences between sequential frames, only limited work hasbeen done for analyzing still images. In this paper, we presentan ROI search algorithm for still images based on a descriptionof the objects in the image. The proposed approach uses no matchingtechniques, but rather uses a set descriptor of the pattern.Experimental results obtained by applying this technique to facialimages show that the proposed methodology is robust with regardto variations in scale and angle as well as stray variationsin the image.