Abstract
The multiplication of publicly available datasets makes it possible to develop Deep Learning models for many real-world applications. However, some domains are still poorly explored, and their related datasets are often small or inconsistent. In addition, some biases linked to the dataset construction or labeling may give the impression that a model is particularly efficient. Therefore, evaluating a model requires a clear understanding of the database. Moreover, a model often reflects a given dataset’s performance and may deteriorate if a shift exists between the training dataset and real-world data.
In this paper, we derive a more consistent and balanced version of the TrashCan [6] image dataset, called UNO, to evaluate models for detecting non-natural objects in the underwater environment. We propose a method to balance the number of annotations and images for cross-evaluation. We then compare the performance of a SOTA object detection model when using TrashCAN and UNO datasets. Additionally, we assess covariate shift by testing the model on an image dataset for real-world application. Experimental results show significantly better and more consistent performance using the UNO dataset.
The UNO database and the code are publicly available at:
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
JAMSTEC: Japan Agency for Marine Earth Science and Technology. https://www.jamstec.go.jp/e/
Canals, M., et al.: The quest for seafloor Macrolitter: a critical review of background knowledge, current methods and future prospects. Environ. Res. Lett., November 2020. https://doi.org/10.1088/1748-9326/abc6d4. publisher: IOP Publishing
Ferrera, M., Creuze, V., Moras, J., Trouvé-Peloux, P.: AQUALOC: an underwater dataset for visual-inertial-pressure localization. Int. J. Rob. Res. 38(14), 1549–1559 (2019). https://doi.org/10.1177/0278364919883346
Fulton, M., Hong, J., Islam, M.J., Sattar, J.: Robotic detection of marine litter using deep visual detection models. In: 2019 International Conference on Robotics and Automation (ICRA) (2019). https://doi.org/10.1109/ICRA.2019.8793975
Haefliger, H.: Deepseawaste (2019). https://www.kaggle.com/henryhaefliger/deepseawaste
Hong, J., Fulton, M., Sattar, J.: A generative approach towards improved robotic detection of marine litter. In: 2020 IEEE International Conference on Robotics and Automation (ICRA) (2020)
Hong, J., Michael, F., Sattar, J.: TrashCan: a semantically-segmented dataset towards visual detection of marine debris. arXiv (2020). https://conservancy.umn.edu/handle/11299/214865
Hu, Y., Pateux, S., Gripon, V.: Squeezing backbone feature distributions to the max for efficient few-shot learning. Algorithms (2022). https://doi.org/10.3390/a15050147, https://hal.archives-ouvertes.fr/hal-03675145
Jocher, G., et al.: ultralytics/yolov5: v6.0 - yolov5 (2021). https://doi.org/10.5281/zenodo.5563715
Madricardo, F., et al.: How to deal with seafloor marine litter: an overview of the state-of-the-art and future perspectives. Front. Mar. Sci. (2020). https://doi.org/10.3389/fmars.2020.505134, https://www.frontiersin.org/article/10.3389/fmars.2020.505134
Ren, S., He, K., Girshick, R., Sun, J.: Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans. Pattern Anal. Mach. Intell. (2015). https://doi.org/10.1109/TPAMI.2016.2577031
Schwerin, P., Wäscher, G.: The bin-packing problem: A problem generator and some numerical experiments with FFD packing and MTP. Int. Trans. Oper. Res. (2006). https://doi.org/10.1111/j.1475-3995.1997.tb00093.x
Tan, M., Le, Q.: EfficientNet: Rethinking model scaling for convolutional neural networks. In: Chaudhuri, K., Salakhutdinov, R. (eds.) Proceedings of the 36th International Conference on Machine Learning. Proceedings of Machine Learning Research, vol. 97. PMLR (2019). https://proceedings.mlr.press/v97/tan19a.html
Tzeng, E., Hoffman, J., Saenko, K., Darrell, T.: Adversarial discriminative domain adaptation. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE Computer Society, Los Alamitos, CA, USA (2017). https://doi.org/10.1109/CVPR.2017.316, https://doi.ieeecomputersociety.org/10.1109/CVPR.2017.316
Veit, A., Matera, T., Neumann, L., Matas, J., Belongie, S.: Coco-text: dataset and benchmark for text detection and recognition in natural images (2016), https://arxiv.org/abs/1601.07140
Wu, H., Xin, M., Fang, W., Hu, H.M., Hu, Z.: Multi-level feature network with multi-loss for person re-identification. IEEE Access (2019). https://doi.org/10.1109/ACCESS.2019.2927052
Zhang, H., Cao, Z., Yan, Z., Zhang, C.: Sill-net: feature augmentation with separated illumination representation (2021)
Acknowledgements
This research has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 101000832 (Maelstrom project).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 Springer Nature Switzerland AG
About this paper
Cite this paper
Barrelet, C., Chaumont, M., Subsol, G., Creuze, V., Gouttefarde, M. (2023). From TrashCan to UNO: Deriving an Underwater Image Dataset to Get a More Consistent and Balanced Version. In: Rousseau, JJ., Kapralos, B. (eds) Pattern Recognition, Computer Vision, and Image Processing. ICPR 2022 International Workshops and Challenges. ICPR 2022. Lecture Notes in Computer Science, vol 13645. Springer, Cham. https://doi.org/10.1007/978-3-031-37731-0_30
Download citation
DOI: https://doi.org/10.1007/978-3-031-37731-0_30
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-37730-3
Online ISBN: 978-3-031-37731-0
eBook Packages: Computer ScienceComputer Science (R0)
