TASK-AWARE SPATIAL CRITICALITY ASSESSMENT OF CAMERA LENS DEFECTS WITH INCOMPLETE MASK EVIDENCE
DOI:
https://doi.org/10.5281/zenodo.22980855Abstract
Camera lens defects may cover the same number of pixels yet obstruct very different parts of a useful viewing region. This paper separates geometric defect coverage from task-aware spatial criticality. It also handles incomplete masks by reporting deterministic lower and upper bounds. A maximum regional score prevents serious obstruction in a small critical region from being hidden by aggregation. Computations on fixed masks show the effects of location, regional dilution, and unresolved pixels. An exhaustive test of 196,608 scalar bounds found no violations under the stated assumptions. The method is an assessment layer for segmentation output, not a claim of calibrated safety risk.Keywords
camera lens defects, spatial criticality, incomplete masks, region of interest, interval assessment, visual reliabilityReferences
· Das, A., Krížek, P., Sistu, G., Bürger, F., Madasamy, S., Uricár, M., Kumar, V. R., & Yogamani, S. (2020). TiledSoilingNet: Tile-level soiling detection on automotive surround-view cameras using coverage metric. In 2020 IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC) (pp. 1–6). IEEE. DOI: 10.1109/ ITSC45102.2020.9294677. IEEE/DOI record
· Yogamani, S., Hughes, C., Horgan, J., Sistu, G., Varley, P., O’Dea, D., Uricár, M., Milz, S., Simon, M., Amende, K., Witt, C., Rashed, H., Chennupati, S., Nayak, S., Mansoor, S., Perrotton, X., & Pérez, P. (2019). WoodScape: A multitask, multi-camera fisheye dataset for autonomous driving. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (pp. 9308–9318). DOI: 10.1109/ICCV.2019.00940. Official CVF publication
· Ronneberger, O., Fischer, P., & Brox, T. (2015). U-Net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, Lecture Notes in Computer Science, Vol. 9351 (pp. 234–241). Springer. DOI: 10.1007/978-3-319-24574-4_28. Springer publication
· Axmedov, A. G. (2026). Uncertainty guided classifier gated segmentation for efficient camera lens defect localization. Manuscript submitted for publication.
· Axmedov, A. G. (2026). Valid-mask temporal consistency filtering for persistent camera lens defect localization and shadow suppression. Manuscript submitted for publication.
· Yang, F., Duan, X., Li, F., & Zhang, L. (2026). A static-to-temporal framework for interpretable camera lens soiling severity estimation in autonomous driving. Sensors, 26(11), 3533. DOI: 10.3390/s26113533. Sensors article
· Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017). On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning, Proceedings of Machine Learning Research, 70, 1321–1330. Official PMLR publication
· Geifman, Y., & El-Yaniv, R. (2019). SelectiveNet: A deep neural network with an integrated reject option. In Proceedings of the 36th International Conference on Machine Learning, Proceedings of Machine Learning Research, 97, 2151–2159. Official PMLR publication
· Beránek, F., Diviš, V., & Gruber, I. (2025). Soiling detection for advanced driver assistance systems. In Seventeenth International Conference on Machine Vision (ICMV 2024). SPIE. DOI: 10.1117/12.3055223. Institutional publication record
· International Organization for Standardization. (2022). ISO 21448:2022 Road vehicles — Safety of the intended functionality. ISO. Official ISO 21448:2022 page
Downloads
Published
How to Cite
Issue
Section
