EUVIMEDEUVIMED entdecken
Crossref

Revisiting Gradient-Based Uncertainty for Monocular Depth Estimation

Julia Hornauer, Amir El-Ghoussani, Vasileios Belagiannis

IEEE Transactions on Pattern Analysis and Machine Intelligence · 2025-06 · 47 (6) · 4395-4408

DOI: 10.1109/tpami.2025.3541964 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025-06. Verfasst von Julia Hornauer und 2 weiteren Autor:innen. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 52 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Monocular depth estimation, similar to other image-based tasks, is prone to erroneous predictions due to ambiguities in the image, for example, caused by dynamic objects or shadows. For this reason, pixel-wise uncertainty assessment is required for safety-critical applications to highlight the areas where the prediction is unreliable. We address this in a post hoc manner and introduce gradient-based uncertainty estimation for already trained depth estimation models. To extract gradients without depending on the ground truth depth, we introduce an auxiliary loss function based on the consistency of the predicted depth and a reference depth. The reference depth, which acts as pseudo ground truth, is in fact generated using a simple image or feature augmentation, making our approach simple and effective. To obtain the final uncertainty score, the derivatives w.r.t. the feature maps from single or multiple layers are calculated using back-propagation. We demonstrate that our gradient-based approach is effective in determining the uncertainty without re-training using the two standard depth estimation benchmarks KITTI and NYU. In particular, for models trained with monocular sequences and therefore most prone to uncertainty, our method outperforms related approaches.

Zitieren

APA

Hornauer, J., El-Ghoussani, A., & Belagiannis, V. (2025). Revisiting Gradient-Based Uncertainty for Monocular Depth Estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(6), 4395-4408. https://doi.org/10.1109/tpami.2025.3541964

Vancouver

Hornauer J, El-Ghoussani A, Belagiannis V. Revisiting Gradient-Based Uncertainty for Monocular Depth Estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2025;47(6):4395-4408. doi:10.1109/tpami.2025.3541964

BibTeX
@article{Hornauer2025,
  author = {Hornauer, Julia and El-Ghoussani, Amir and Belagiannis, Vasileios},
  title = {Revisiting Gradient-Based Uncertainty for Monocular Depth Estimation},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2025},
  volume = {47},
  number = {6},
  pages = {4395--4408},
  doi = {10.1109/tpami.2025.3541964}
}
Kostenlos starten →