Abstract: Two computer-assisted photo-identification methods for sperm whales (Physeter macrocephalus), namely the Highlight method (Whitehead, 1990) and the Europhlukes method (based on Huele et al., 2000), were compared. Performance was measured in terms of speed and accuracy. A test set was constructed containing two photographs of each of 296 individuals. The test set was divided into three classes of photographic quality and three classes of pattern distinctiveness. Both programs met requirements for rapid matching; the mean extraction times were 74.2 and 90.1 s per image for the Highlight and the Europhlukes methods, respectively. The two methods performed similarly with respect to accuracy. Accuracy improved by using higher-quality photographs or photographs representing more distinctive flukes. Still, even when using only the higher-quality photographs, 12.4% of the matches were not included in the top nine of the list of potential matches by the Highlight method compared to 14.0% for the Europhlukes method. The rate of failure to find the true match in the top nine was only 3.3% when both methods were used together, however. It is, therefore, recommended that for improved matching, both methods should be used in tandem or that an integrated program, which combines the two methods, should be developed.
Key Words: SPERM WHALE; PHYSETER MACROCEPHALUS; PHOTO-IDENTIFICATION; COMPUTER-ASSISTED MATCHING; HIGHLIGHT METHOD; EUROPHLUKES METHOD
Document Type: Research article
Page Numbers: 243 - 247