TY - GEN
T1 - Hammerhead Shark Species Monitoring with Deep Learning
AU - Peña, Alvaro
AU - Pérez, Noel
AU - Benítez, Diego S.
AU - Hearn, Alex
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - In this paper, we propose a new automated method based on deep convolutional neural networks to detect and track critically endangered hammerhead sharks in video sequences. The proposed method improved the standard YOLOv3 deep architecture by adding 18 more layers (16 convolutional and 2 Yolo layers), which increased the model performance in detecting the species under analysis at different scales. According to the frame analysis based validation, the proposed method outperformed the standard YOLOv3 model and was similar to the mask R-CNN model in terms of accuracy scores for the majority of inspected frames. Also, the mean of precision and recall on an experimental frames dataset formed using the 10-fold cross-validation method highlighted that the proposed method outperformed the remaining architectures, reaching scores of 0.99 and 0.93, respectively. Furthermore, the methods were able to avoid introducing false positive detection. However, they were unable to handle the problem of species occlusion. Our results indicate that the proposed method is a feasible alternative tool that could help to monitor relative abundance of hammerhead sharks in the wild.
AB - In this paper, we propose a new automated method based on deep convolutional neural networks to detect and track critically endangered hammerhead sharks in video sequences. The proposed method improved the standard YOLOv3 deep architecture by adding 18 more layers (16 convolutional and 2 Yolo layers), which increased the model performance in detecting the species under analysis at different scales. According to the frame analysis based validation, the proposed method outperformed the standard YOLOv3 model and was similar to the mask R-CNN model in terms of accuracy scores for the majority of inspected frames. Also, the mean of precision and recall on an experimental frames dataset formed using the 10-fold cross-validation method highlighted that the proposed method outperformed the remaining architectures, reaching scores of 0.99 and 0.93, respectively. Furthermore, the methods were able to avoid introducing false positive detection. However, they were unable to handle the problem of species occlusion. Our results indicate that the proposed method is a feasible alternative tool that could help to monitor relative abundance of hammerhead sharks in the wild.
KW - Deep convolutional neural networks
KW - Hammerhead shark detection and tracking
KW - Mask R-CNN architecture
KW - Real-time detector
KW - YOLOv3 architecture
UR - https://www.scopus.com/pages/publications/85103296578
U2 - 10.1007/978-3-030-69774-7_4
DO - 10.1007/978-3-030-69774-7_4
M3 - Contribución a la conferencia
AN - SCOPUS:85103296578
SN - 9783030697730
T3 - Communications in Computer and Information Science
SP - 45
EP - 59
BT - Applications of Computational Intelligence - 3rd IEEE Colombian Conference, ColCACI 2020, Revised Selected Papers
A2 - Orjuela-Cañón, Alvaro David
A2 - Lopez, Jesus
A2 - Arias-Londoño, Julián David
A2 - Figueroa-García, Juan Carlos
PB - Springer Science and Business Media Deutschland GmbH
T2 - 3rd IEEE Colombian Conference on Applications of Computational Intelligence, IEEE ColCACI 2020
Y2 - 7 August 2020 through 8 August 2020
ER -