Abstract
Object localization is a fundamental and important task in computer vision, that is used as a pre-processing step for object detection and semantic segmentation. However, fully supervised object localization requires bounding boxes and pixel-level labels, and these annotations are expensive. For this reason, Weakly Supervised Object Localization (WSOL) with image-level (weak) supervision has been the focus of much research in recent years. However, WSOL requires a large dataset to detect the region of an object in images with high performance. When the large dataset is unavailable, it is difficult to localize the image with high performance. This paper proposes a method for extracting target regions using small amounts of target and background images with image-level labels. The proposed method enables the detection of object locations with high performance using relatively less training images by classifying multiple patches cut from the image. This object localization method differs from the typical WSOL method that takes a single image as input and detects the location of an object because it assumes a small patch of area as input. The label of the patch cropped from the image must be labeled with the ground truth. However, the proposed method uses labels attached to images because ground truth labeling is costly. Instead, in the proposed method, the network learns by learning many “background” labeled background patches, and learns to induce the network to classify the mislabeled background patches that resemble ground truth as background. We call this key idea Decision-Boundary Induction(DBI). Moreover, learning many background patches for such a DBI is what we call background learning. In our experiments, we verified that decision boundaries are induced, and accordingly, we could roughly extract the target region. Also, we showed that the Loc. Acc. is higher than that of WSOL.
Prof. Masaru Tanaka deceased in June 2021.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Wang, C., Ren, W., Huang, K., Tan, T.: Weakly supervised object localization with latent category learning. In: Fleet, D., Pajdla, T., Schiele, B., Tuytelaars, T. (eds.) ECCV 2014. LNCS, vol. 8694, pp. 431–445. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-10599-4_28
Song, H.O., Girshick, R., Jegelka, S., Mairal, J., Harchaoui, Z., Darrell, T.: On learning to localize objects with minimal supervision. In: Proceedings of the 31st International Conference on International Conference on Machine Learning, ICML 2014, vol. 32, pp. 1611–1619. JMLR.org (2014)
Cinbis, R., Verbeek, J., Schmid, C.: Multi-fold MIL Training for weakly supervised object localization (2014). https://doi.org/10.1109/CVPR.2014.309
Oquab, M., Bottou, L., Laptev, I., Sivic, J.: Is object localization for free? - weakly-supervised learning with convolutional neural networks (2015). https://doi.org/10.1109/CVPR.2015.7298668
Liang, X., Liu, S., Wei, Y., Liu, L., Lin, L., Yan, S.: Towards computational baby learning: a weakly-supervised approach for object detection (2015). https://doi.org/10.1109/ICCV.2015.120
Teh, E.W., Rochan, M., Wang, Y.: Attention networks for weakly supervised object localization. In: Wilson, R.C., Hancock, E.R., Smith, W.A.P. (eds.) Proceedings of the British Machine Vision Conference (BMVC), pp. 52–15211. BMVA Press (2016). https://doi.org/10.5244/C.30.52
Simonyan, K., Vedaldi, A., Zisserman, A.: Deep inside convolutional networks: Visualising image classification models and saliency maps. CoRR abs/1312.6034 (2013)
Pathak, D., Krahenbuhl, P., Darrell, T.: Constrained convolutional neural networks for weakly supervised segmentation. In: ICCV, pp. 1796–1804 (2015)
Kolesnikov, A., Lampert, C.H.: Seed, expand and constrain: three principles for weakly-supervised image segmentation. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016. LNCS, vol. 9908, pp. 695–711. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46493-0_42
Khoreva, A., Benenson, R., Omran, M., Hein, M., Schiele, B.: Weakly supervised object boundaries. In: CVPR, pp. 183–192 (2016)
Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., Torralba, A.: Learning deep features for discriminative localization (2016). https://doi.org/10.1109/CVPR.2016.319
Choe, J., Lee, S., Shim, H.: Attention-based dropout layer for weakly supervised single object localization and semantic segmentation. IEEE Trans. Pattern Anal. Mach. Intell. 43(12), 4256–4271 (2021). https://doi.org/10.1109/TPAMI.2020.2999099
Wei, Y., Feng, J., Liang, X., Cheng, M.-M., Zhao, Y., Yan, S.: Object region mining with adversarial erasing: a simple classification to semantic segmentation approach (2017). https://doi.org/10.1109/CVPR.2017.687
Bae, W., Noh, J., Kim, G.: Rethinking class activation mapping for weakly supervised object localization. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M. (eds.) ECCV 2020. LNCS, vol. 12360, pp. 618–634. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-58555-6_37
Zhang, C.-L., Cao, Y.-H., Wu, J.: Rethinking the route towards weakly supervised object localization (2020). https://doi.org/10.1109/CVPR42600.2020.01347
Yun, S., Han, D., Chun, S., Oh, S.J., Yoo, Y., Choe, J.: CutMix: regularization strategy to train strong classifiers with localizable features (2019). https://doi.org/10.1109/ICCV.2019.00612
Singh, K.K., Lee, Y.J.: Hide-and-seek: forcing a network to be meticulous for weakly-supervised object and action localization (2017). https://doi.org/10.1109/ICCV.2017.381
Rahimi, A., Shaban, A., Ajanthan, T., Hartley, R., Boots, B.: Pairwise similarity knowledge transfer for weakly supervised object localization. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M. (eds.) ECCV 2020. LNCS, vol. 12369, pp. 395–412. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-58586-0_24
Hwang, S., Kim, H.-E.: Self-transfer learning for weakly supervised lesion localization. In: Ourselin, S., Joskowicz, L., Sabuncu, M.R., Unal, G., Wells, W. (eds.) MICCAI 2016. LNCS, vol. 9901, pp. 239–246. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46723-8_28
Chawla, N., Japkowicz, N., Kołcz, A.: Editorial: special issue on learning from imbalanced data sets. SIGKDD Explor. 6, 1–6 (2004). https://doi.org/10.1145/1007730.1007733
Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. In: ICLR (2015)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: CVPR (2016)
Acknowledgments
I am grateful to Associate Professor Takafumi Amaha of Fukuoka University for advice on writing the paper. I would like to take this opportunity to thank him.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Nakamura, R., Ueda, Y., Tanaka, M., Fujiki, J. (2023). Rough Target Region Extraction with Background Learning. In: Na, I., Irie, G. (eds) Frontiers of Computer Vision. IW-FCV 2023. Communications in Computer and Information Science, vol 1857. Springer, Singapore. https://doi.org/10.1007/978-981-99-4914-4_3
Download citation
DOI: https://doi.org/10.1007/978-981-99-4914-4_3
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-99-4913-7
Online ISBN: 978-981-99-4914-4
eBook Packages: Computer ScienceComputer Science (R0)