1. Photo-identification (photo-id) is a non-invasive capture-recapture method used by marine researchers to monitor cetacean populations.
2. This paper presents a fully automatic photo-id aid capable of providing most likely matches based on all available information without the need for data pre-processing such as cropping.
3. The system was evaluated against multiple real-life photo-id catalogues, achieving high accuracy for both dorsal fin detection and individual classification tasks.
The article is generally reliable and trustworthy, as it provides evidence to support its claims and presents the results of its evaluation in an objective manner. The authors have also taken into account potential biases in their research, such as the difficulty of accurately detecting cetaceans in unedited field imagery, and have attempted to mitigate these through post-processing techniques. However, there are some areas where the article could be improved upon. For example, while the authors note that their system can handle previously uncatalogued individuals, they do not provide any evidence or further explanation as to how this is achieved. Additionally, while the authors mention that their system can detect cetaceans in unedited field imagery, they do not provide any details on how this is done or what techniques are used to achieve this goal. Finally, while the authors present results from two different catalogues (Tanzania and UK), they do not provide any comparison between them or discuss any potential differences between them which could affect the accuracy of their system.