The Salamandra infraimmaculata, commonly known as the Near Eastern fire salamander, has been assigned the conservation status of “Near Threatened” on the IUCN Red List of Threatened Species. The Society for the Protection of Nature in Israel, in collaboration with Yarok Balev NGO, works to track Salamander population size by identifying individual salamanders by human experts based on digital images collected during annual winter monitoring projects. While this process is currently being done manually, we propose the development of an automated algorithm to process the images so that the identification of individual salamanders can take place without human intervention.

The main idea behind our approach is to treat the spot pattern on the salamanders as a type of identifying fingerprint. To implement our algorithm, we developed a multi-step process that involves several fundamental procedures. Initially, we perform image segmentation, which enables us to identify the salamander and its unique spots.

We generate a straightened image of the salamander using SVD and the Dijkstra algorithm to transform the pixels of the salamander, preserving the distance of each pixel from the head of the salamander and its skeleton. Then spatial and geometric features from the salamander’s spots are gathered, including size, shape, and center position and we construct a function to estimate the location of matching spots based on geometric features. We then used a neural network to determine the level of match between spots from different images. To do this we used a triplet loss function, training the net with examples of matching spots of the same salamander and non matching spots. Finally we use the output of the neural network to establish whether the salamander is the same or different according to how well the spot pattern fits previous images of salamanders.