| Summary: | Humpback whales (Megaptera novaeangliae) are characterized by variable natural pigmentation patterns and scarring marks on their tail flukes and other regions of the body. The ability to recognize individual humpback whales from photographs of their tail fluke pigmentation and scarring patterns was first realized by researchers in the 1970s. Since that time, the technique of photo-identification has been widely used on humpback whale populations around the world to determine many aspects of their biology, ecology, and behavior. However, a significant and growing problem exists with comparing photo-identification data sets. Traditional methods of matching photographs of ventral fluke surfaces require manual pairwise comparison of all images within and among data sets, which are often very large. This process requires substantial time, effort, and expertise. Furthermore, as each of the data sets grows, the number of comparisons required increases exponentially. Computer-based matching systems have been developed for other marine species, including some cetaceans. However, attempts to develop a computerized system for humpback whales are still in their infancy and have focused largely on patch distribution of black and white markings on the fluke. To date, none of the systems used for computer-aided fluke identification employ the wide range of features and unique properties of humpback whale flukes that can be utilized to more efficiently identify individuals.
|
|---|