Abstract

Individual fish identification and recognition is an important step in the conservation and management of fisheries. One of most frequently used methods involves capturing and tagging fish. However, these processes have been reported to cause tissue damage, premature tag loss, and decreased swimming capacity. More recently, marine video recordings have been extensively used for monitoring fish populations. However, these require visual inspection to identify individual fish. In this work, we proposed an automatic method for the identification of individual brown trouts, Salmo trutta. We developed a deep convolutional architecture for this purpose. Specifically, given two fish images, multi-scale convolutional features were extracted to capture low-level features and high-level semantic components for embedding space representation. The extracted features were compared at each scale for capturing representation for individual fish identification. The method was evaluated on a dataset called NINA204 based on 204 videos of brown trout and on a dataset TROUT39 containing 39 brown trouts in 288 frames. The identification method distinguished individual fish with 94.6% precision and 74.3% recall on a NINA204 video sequence with significant appearance and shape variation. The identification method takes individual fish and is able to distinguish them with precision and recall percentages of 94.6% and 74.3% on NINA204 for a video sequence with significant appearance and shape variation.