FishSizer: Software solution for efficiently measuring larval fish size.

Length and depth of fish larvae are part of the fundamental measurements in many marine ecology studies involving early fish life history. Until now, obtaining these measurements has required intensive manual labor and the risk of inter- and intra-observer variability.We developed an open-source software solution to semi-automate the measurement process and thereby reduce both time consumption and technical variability. Using contrast-based edge detection, the software segments images of a fish larva into "larva" and "background." Length and depth are extracted from the "larva" segmentation while taking curvature of the larva into consideration. The graphical user interface optimizes workflow and ease of usage, thereby reducing time consumption for both training and analysis. The software allows for visual verification of all measurements.A comparison of measurement methods on a set of larva images showed that this software reduces measurement time by 66%-78% relative to commonly used software.Using this software instead of the commonly used manual approach has the potential to save researchers from many hours of monotonous work. No adjustment was necessary for 89% of the images regarding length (70% for depth). Hence, the only workload on most images was the visual inspection. As the visual inspection and manual dimension extraction works in the same way as currently used software, we expect no loss in accuracy.

Automatic detection and classification of baleen whale social calls using convolutional neural networks.

Passive acoustic monitoring has proven to be an indispensable tool for many aspects of baleen whale research. Manual detection of whale calls on these large data sets demands extensive manual labor. Automated whale call detectors offer a more efficient approach and have been developed for many species and call types. However, calls with a large level of variability such as fin whale (Balaenoptera physalus) 40 Hz call and blue whale (B. musculus) D call have been challenging to detect automatically and hence no practical automated detector exists for these two call types. Using a modular approach consisting of faster region-based convolutional neural network followed by a convolutional neural network, we have created automated detectors for 40 Hz calls and D calls. Both detectors were tested on recordings with high- and low density of calls and, when selecting for detections with high classification scores, they were shown to have precision ranging from 54% to 57% with recall ranging from 72% to 78% for 40 Hz and precision ranging from 62% to 64% with recall ranging from 70 to 73% for D calls. As these two call types are produced by both sexes, using them in long-term studies would remove sex-bias in estimates of temporal presence and movement patterns.