Comparative Study of Lightweight YOLOv12 Models for Real-Time Underwater Object Detection
DOI:
https://doi.org/10.61769/telematika.v20i2.799Keywords:
underwater, object detection, convolutional neural network, efficient model, lightweight YOLOv12Abstract
Deep learning methods in computer vision play a crucial role in object localization using camera-based sensors, with Convolutional Neural Networks serving as the dominant approach for object detection. However, many existing models incur high computational costs due to deep architectures and complex operations, limiting their use for real-time deployment on low-cost, resource-constrained devices. The YOLOv12 architecture offers lightweight variants to improve computational efficiency. This study evaluates the trade-off between efficiency and detection performance by comparing model variants using the number of parameters, floating-point operations, and inference speed, while detection accuracy is measured using mean average precision. The results assess the suitability of lightweight models for real-time deployment in resource-constrained environments such as underwater monitoring and conservation. Experimental results on the Real-World Underwater Object Detection dataset demonstrate that YOLOv12-nano achieves 5.7% lower accuracy compared to YOLOv12-medium but requires only 2.57 million parameters and 6.5 GFLOPs, significantly less than YOLOv12-medium with 20.1 million parameters and 67.8 GFLOPs. Moreover, YOLOv12-small requires 9.26 million parameters and 21.5 GFLOPs, positioning it between nano and medium in terms of complexity while still maintaining competitive accuracy. In the inference process, YOLOv12-nano achieves 16.48 FPS on a 12th Gen Intel(R) Core (TM) i5-12450HX CPU. In comparison, YOLOv12-small runs at 6.28 FPS, while YOLOv12-medium runs at 2.36 FPS. These results indicate that YOLOv12-nano is the most suitable variant for real-time deployment on CPU-based platforms.
References
A. Apprill et al., “Toward a new era of coral reef monitoring,” Environmental Science & Technology, vol. 57, no. 13, pp. 5117–5124, Apr. 2023, doi: 10.1021/acs.est.2c05369.
F. Wu, Z. Cai, S. Fan, R. Song, L. Wang, and W. Cai, “Fish target detection in underwater blurred scenes based on improved YOLOv5,” IEEE Access, vol. 11, pp. 122911–122925, 2023, doi: 10.1109/ACCESS.2023.3328940.
T.-N. Pham, V.-H. Nguyen, K.-R. Kwon, J.-H. Kim, and J.-H. Huh, “Improved YOLOv5-based deep learning system for jellyfish detection,” IEEE Access, vol. 12, pp. 87838–87849, 2024, doi: 10.1109/ACCESS.2024.3405452.
P. Pachaiyappan, G. Chidambaram, A. Jahid, and M. H. Alsharif, “Enhancing underwater object detection and classification using advanced imaging techniques: A novel approach with diffusion models,” Sustainability, vol. 16, no. 17, Art. no. 7488, 2024, doi: 10.3390/su16177488.
D. Song and H. Huo, “Lightweight underwater target detection algorithm based on YOLOv8n,” Electronics, vol. 14, no. 9, Art. no. 1810, 2025, doi: 10.3390/electronics14091810.
Y. Xu and X. Xiao, “Exploring the depth from EAST: Efficient aggregated state-space tanh-tuned model for underwater object detection,” IEEE Signal Processing Letters, vol. 32, pp. 3809–3813, 2025, doi: 10.1109/LSP.2025.3606841.
Z. Song, X. Zhang, and P. Tan, “YOLO-Fast: A lightweight object detection model for edge devices,” The Journal of Supercomputing, vol. 81, no. 5, Apr. 2025, doi: 10.1007/s11227-025-07172-3.
J. Lei, H. Wang, Z. Lei, J. Li, and S. Rong, “CNN–Transformer hybrid architecture for underwater sonar image segmentation,” Remote Sensing, vol. 17, no. 4, Art. no. 707, 2025, doi: 10.3390/rs17040707.
J. Liu, S. Liu, S. Xu, and C. Zhou, “Two-stage underwater object detection network using Swin Transformer,” IEEE Access, vol. 10, pp. 117235–117247, 2022, doi: 10.1109/ACCESS.2022.3219592.
L. Guo, X. Liu, D. Ye, X. He, J. Xia, and W. Song, “Underwater object detection algorithm integrating image enhancement and deformable convolution,” Ecological Informatics, vol. 89, Art. no. 103185, 2025, doi: 10.1016/j.ecoinf.2025.103185.
X. Qin, C. Yu, B. Liu, and Z. Zhang, “YOLOv8-FASG: A high-accuracy fish identification method for underwater robotic system,” IEEE Access, vol. 12, pp. 73354–73362, 2024, doi: 10.1109/ACCESS.2024.3404867.
W. Yi, J. Yang, and L. Yan, “Research on underwater small target detection technology based on single-stage USSTD-YOLOv8n,” IEEE Access, vol. 12, pp. 69633–69641, 2024, doi: 10.1109/ACCESS.2024.3400962.
J. Huang, C. Fang, X. Zheng, and J. Liu, “YOLOv8-UC: An improved YOLOv8-based underwater object detection algorithm,” IEEE Access, vol. 12, pp. 172186–172195, 2024, doi: 10.1109/ACCESS.2024.3496925.
Z. Li, H. Xie, J. Feng, Z. Wang, and Z. Yuan, “YOLOv7-PE: A precise and efficient enhancement of YOLOv7 for underwater target detection,” IEEE Access, vol. 12, pp. 133937–133951, 2024, doi: 10.1109/ACCESS.2024.3417322.
Y. Tian, Q. Ye, and D. Doermann, “YOLOv12: Attention-centric real-time object detectors,” arXiv preprint arXiv:2502.12524, Feb. 2025. [Online]. Available: http://arxiv.org/abs/2502.12524
S. Liu, L. Qi, H. Qin, J. Shi, and J. Jia, “Path aggregation network for instance segmentation,” in 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 8759–8768, doi: 10.1109/CVPR.2018.00913.
X. Li et al., “Generalized focal loss: Learning qualified and distributed bounding boxes for dense object detection,” arXiv preprint arXiv:2006.04388, Jun. 2020. [Online]. Available: http://arxiv.org/abs/2006.04388
Z. Zheng et al., “Enhancing geometric factors in model learning and inference for object detection and instance segmentation,” IEEE Transactions on Cybernetics, vol. 52, no. 8, pp. 8574–8586, 2022, doi: 10.1109/TCYB.2021.3095305.
A. U. Ruby, “Binary cross entropy with deep learning technique for image classification,” International Journal of Advanced Trends in Computer Science and Engineering, vol. 9, no. 4, pp. 5393–5397, Aug. 2020, doi: 10.30534/ijatcse/2020/175942020.
N. Shahadat and A. S. Maida, “Analyzing parameter-efficient convolutional neural network architectures for visual classification,” Sensors, vol. 25, no. 24, Art. no. 7663, Dec. 2025, doi: 10.3390/s25247663.
C. Fu et al., “Rethinking general underwater object detection: Datasets, challenges, and solutions,” Neurocomputing, vol. 517, pp. 243–256, 2023, doi: 10.1016/j.neucom.2022.10.039.
A. Wang et al., “YOLOv10: Real-time end-to-end object detection,” arXiv preprint arXiv:2405.14458, Oct. 2024. [Online]. Available: http://arxiv.org/abs/2405.14458
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Hebron Prasetya, Revin R. Balo, Tasya Tumbal, Alwin M. Sambul, Muhamad Dwisnanto Putro
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
You are free to:
- Share — copy and redistribute the material in any medium or format for any purpose, even commercially.
- Adapt — remix, transform, and build upon the material for any purpose, even commercially.
- The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms:
- Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
- ShareAlike — If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.
- No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
Notices:
You do not have to comply with the license for elements of the material in the public domain or where your use is permitted by an applicable exception or limitation.
No warranties are given. The license may not give you all of the permissions necessary for your intended use. For example, other rights such as publicity, privacy, or moral rights may limit how you use the material.
