Photo-identification of humpback and gray whales using affine moment invariants

Authors:
N. Kehtarnavaz;V. Peddigari;C. Chandan;W. Syed;G. Hillman;B. Wursig
Affiliations:
Department of Electrical Engineering, University of Texas at Dallas;Department of Electrical Engineering, University of Texas at Dallas;Department of Electrical Engineering, University of Texas at Dallas;Department of Electrical Engineering, University of Texas at Dallas;Department of Pharmacology, University of Texas Medical Branch at Galveston;Marine Biology Department, Texas A&M University
Venue:
SCIA'03 Proceedings of the 13th Scandinavian conference on Image analysis
Year:
2003

Citing 1
Cited 1

Interactive segmentation with Intelligent Scissors

Graphical Models and Image Processing

Affine invariant comparison of point-sets using convex hulls and hausdorff distances

Pattern Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper discusses the identification of humpback and gray whales based on images of their flukes. White patches appearing on a fluke constitute unique identifying features which are used here to match a query fluke image against a database of fluke images. The fluke area is first extracted from a fluke image via the live-wire edge detection algorithm. Optimal thresholding is then applied to the fluke area to obtain the patches on the fluke automatically. Next, the affine moment invariants of patches are computed and stored in a feature database. Finally, the database images are ranked depending on the degree of similarity between the query and database invariants. The results obtained are presented in the form of a performance box plot showing the reduction in the amount of manual search for various percentages of test images.