A Comparison of Deep Learning Approach for Underwater Object Detection

Nurcahyani Wulandari; Igi Ardiyanto; Hanung Adi Nugroho

doi:10.29207/resti.v6i2.3931

Nurcahyani Wulandari Universitas Gadjah Mada
Igi Ardiyanto Universitas Gadjah Mada
Hanung Adi Nugroho Universitas Gadjah Mada

DOI: https://doi.org/10.29207/resti.v6i2.3931

Keywords: Underwater Object Detection, Faster-RCNN, SSD, RetinaNet, YOLOv3, YOLOv4

Abstract

In recent years, marine ecosystems and fisheries have become potential resources. Therefore, monitoring these objects will be essential to ensure their existence. One of the computer vision techniques is object detection, utilized to recognize and localize objects in underwater scenery. Many studies have been conducted to investigate various deep learning methods implemented in underwater object detection; however, only a few investigations have been performed to compare mainstream object detection algorithms in these circumstances. This article examines various state-of-the-art deep learning methods applied to underwater object detection, including Faster-RCNN, SSD, RetinaNet, YOLOv3, and YOLOv4. We trained five models on the RUIE dataset. The average detection time was used to compare how fast a model can detect an object within an image, and mAP was also applied to measure detection accuracy. All trained models have costs and benefits; SSD was fast but had poor performance; RetinaNet had consistent performance across different thresholds, but the detection speed was slow; YOLOv3 was the fastest and had acceptable performance comparable with RetinaNet; YOLOv4 was good at first, but performance dropped as threshold enlargement; also, YOLOv4 needed extra time to detect objects compared to YOLOv3. There are no models that are fully suited for underwater object detection; nonetheless, when the mAP and average detection time of the five models were compared, we determined that YOLOv3 is the best acceptable model among the evaluated underwater object detection models.

Downloads

Download data is not yet available.

References

Badan Statistik Nasional, “Ekspor Ikan Segar/Dingin Hasil Tangkap menurut Negara Tujuan Utama, 2012-2020,” 2021. https://www.bps.go.id/statictable/2019/02/25/2024/ekspor-ikan-segar-dingin-hasil-tangkap-menurut-negara-tujuan-utama-2012-2020.html (accessed Jan. 12, 2022).

Direktorat Jenderal Penguatan Daya Saing Produk Kelautan dan Perikanan, Statistik Impor Hasil Perikanan Tahun 2016-2020. 2021.

A. A. Sentosa and D. Wijaya, “Potensi invasif ikan zebra Cichlid (Amatitlania nigrofasciata Günther, 1867) di Danau Beratan, Bali ditinjau dari aspek biologinya,” BAWAL Widya Ris. Perikan. Tangkap, vol. 5, no. 2, pp. 113–121, 2013.

M. Dan, A. Ronny, S. Balai, P. Pemulihan, D. Konservasi, and S. Ikan, “Karakteristik Habitat Ikan Kerapu Di Kepulauan Karimunjawa, Jawatengah the Characteristic Habitat of Grouper Fish in Karimunjawa Islands, Central Java,” BAWAL Cilalawi, vol. 7, no. 1, pp. 147–154, 2015.

D. W. H. Tjahjo, S. E. Purnamaningtyas, and A. Suryandari, “Evaluasi Peran Jenis Ikan Dalam Pemanfaatan Sumber Daya Pakan Dan Ruang Di Waduk Ir. H. Djuanda, Jawa Barat,” J. Penelit. Perikan. Indones., vol. 15, no. 4, p. 267, 2017, doi: 10.15578/jppi.15.4.2009.267-276.

Y. X. Bai et al., “Automatic multiple zebrafish tracking based on improved HOG features,” Sci. Rep., vol. 8, no. 1, pp. 1–14, 2018, doi: 10.1038/s41598-018-29185-0.

S. Villon, M. Chaumont, G. Subsol, S. Villéger, T. Claverie, and D. Mouillot, “Coral Reef Fish Detection and Recognition in Underwater Videos by Supervised Machine Learning: Comparison Between Deep Learning and HOG+SVM Methods,” in Lecture Notes in Computer Science, vol. 10016 LNCS, 2016, pp. 160–171.

S. Choi, “Fish identification in underwater video with deep convolutional neural network: SNUMedinfo at LifeCLEF fish task 2015,” CEUR Workshop Proc., vol. 1391, pp. 1–5, 2015.

D. Rathi, S. Jain, and S. Indu, “Underwater Fish Species Classification using Convolutional Neural Network and Deep Learning,” 2017 9th Int. Conf. Adv. Pattern Recognition, ICAPR 2017, pp. 344–349, 2018, doi: 10.1109/ICAPR.2017.8593044.

S. Cui, Y. Zhou, Y. Wang, and L. Zhai, “Fish Detection Using Deep Learning,” Appl. Comput. Intell. Soft Comput., vol. 2020, 2020, doi: 10.1155/2020/3738108.

M. S. Cueto, J. M. B. Diangkinay, K. W. B. Melencion, T. P. Senerado, H. L. P. Taytay, and E. R. E. Tolentino, “Classification of different types of koi fish using convolutional neural network,” Proc. - 5th Int. Conf. Intell. Comput. Control Syst. ICICCS 2021, no. Iciccs, pp. 1135–1142, 2021, doi: 10.1109/ICICCS51141.2021.9432358.

G. Xu et al., “Detection of Bluefin Tuna by Cascade Classifier and Deep Learning for Monitoring Fish Resources,” 2020 Glob. Ocean. 2020 Singapore - U.S. Gulf Coast, pp. 2020–2023, 2020, doi: 10.1109/IEEECONF38699.2020.9389012.

C. S. Arvind, R. Prajwal, P. N. Bhat, A. Sreedevi, and K. N. Prabhudeva, “Fish Detection and Tracking in Pisciculture Environment using Deep Instance Segmentation,” IEEE Reg. 10 Annu. Int. Conf. Proceedings/TENCON, vol. 2019-Octob, pp. 778–783, 2019, doi: 10.1109/TENCON.2019.8929613.

X. Li, M. Shang, H. Qin, and L. Chen, “Fast accurate fish detection and recognition of underwater images with Fast R-CNN,” Ocean. 2015 - MTS/IEEE Washingt., pp. 1–5, 2016, doi: 10.23919/oceans.2015.7404464.

R. Mandal, R. M. Connolly, T. A. Schlacher, and B. Stantic, “Assessing fish abundance from underwater video using deep neural networks,” Proc. Int. Jt. Conf. Neural Networks, vol. 2018-July, pp. 1–6, 2018, doi: 10.1109/IJCNN.2018.8489482.

K. U. Akdemir and E. Alaybeyoglu, “Classification of red mullet, bluefish and haddock caught in the black sea by ‘single shot multibox detection,’” 2021 Int. Conf. Innov. Intell. Syst. Appl. INISTA 2021 - Proc., pp. 21–24, 2021, doi: 10.1109/INISTA52262.2021.9548488.

C. C. Wang, H. Samani, and C. Y. Yang, “Object Detection with Deep Learning for Underwater Environment,” Proc. 4th Int. Conf. Inf. Technol. Res. Bridg. Digit. Divid. Through Multidiscip. Res. ICITR 2019, 2019, doi: 10.1109/ICITR49409.2019.9407797.

M. S. A. Bin Rosli, I. S. Isa, M. I. F. Maruzuki, S. N. Sulaiman, and I. Ahmad, “Underwater Animal Detection Using YOLOV4,” Proc. - 2021 11th IEEE Int. Conf. Control Syst. Comput. Eng. ICCSCE 2021, no. August, pp. 158–163, 2021, doi: 10.1109/ICCSCE52189.2021.9530877.

S. Ren, K. He, R. Girshick, and J. Sun, “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 39, no. 6, pp. 1137–1149, Jun. 2017, doi: 10.1109/TPAMI.2016.2577031.

W. Liu et al., “SSD: Single shot multibox detector,” in Lecture Notes in Computer Science, Dec. 2016, vol. 9905 LNCS, pp. 21–37, doi: 10.1007/978-3-319-46448-0_2.

T. Lin, P. Goyal, R. Girshick, K. He, and P. Dollar, “Focal Loss for Dense Object Detection,” in 2017 IEEE International Conference on Computer Vision (ICCV), Oct. 2017, pp. 2999–3007, doi: 10.1109/ICCV.2017.324.

J. Redmon and A. Farhadi, “YOLOv3: An Incremental Improvement,” Proc. 4th Int. Conf. Inf. Technol. Res. Bridg. Digit. Divid. Through Multidiscip. Res. ICITR 2019, Apr. 2018, [Online]. Available: http://arxiv.org/abs/1804.02767.

A. Bochkovskiy, C. Wang, and H. M. Liao, “YOLOv4: Optimal Speed and Accuracy of Object Detection,” Apr. 2020, [Online]. Available: http://arxiv.org/abs/2004.10934.

R. Liu, X. Fan, M. Zhu, M. Hou, and Z. Luo, “Real-World Underwater Enhancement: Challenges, Benchmarks, and Solutions Under Natural Light,” IEEE Trans. Circuits Syst. Video Technol., vol. 30, no. 12, pp. 4861–4875, Dec. 2020, doi: 10.1109/TCSVT.2019.2963772.

M. Everingham, L. Van Gool, C. K. I. Williams, J. Winn, and A. Zisserman, “The pascal visual object classes (VOC) challenge,” Int. J. Comput. Vis., vol. 88, no. 2, pp. 303–338, 2010, doi: 10.1007/s11263-009-0275-4.