Deep Active Learning for Efficient Training of a LiDAR 3D Object Detector
Open accessAbstract
Training a deep object detector for autonomous driving requires a huge amount of labeled data. While recording data via on-board sensors such as camera or LiDAR is relatively easy, annotating data is very tedious and time-consuming, especially when dealing with 3D LiDAR points or radar data. Active learning has the potential to minimize human annotation efforts while maximizing the object detector's performance. In this work, we propose an active learning method to train a LiDAR 3D object detector with the least amount of labeled training data necessary. The detector leverages 2D region proposals generated from the RGB images to reduce the search space of objects and speed up the learning process. Experiments show that our proposed method works under different uncertainty estimations and query functions, and can save up to 60% of the labeling efforts while reaching the same network performance.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
A significant part of the works found in the literature addresses object detection over 2D data, while 3D data are also focused on by other approaches.
“Even though a significant part of the works found in the literature addresses OD over 2D data (see Table 2), 3D data are also focused on by other approaches (e.g., [15,53]), in which the application of AL techniques proved to be beneficial.”
From Ten Years of Active Learning Techniques and Object Detection: A Systematic Review (2023) · cites this paper as [53]
Uncertainty is estimated using three different classification predictive probabilities: softmax output, Monte-Carlo dropout, and deep ensembles.
“Uncertainty is estimated using three different classification predictive probabilities for comparison purposes: (i) softmax output, (ii) Monte-Carlo dropout, and (iii) deep ensembles.”
From Ten Years of Active Learning Techniques and Object Detection: A Systematic Review (2023) · cites this paper as [53]
A method that combines 3D LiDAR and 2D data is proposed in [53], where bounding boxes obtained by a 2D image detector are used to locate 3D objects.
“A method that combines 3D LiDARand2Ddataisproposed in [53], where the bounding boxes obtained by the 2D image detector are used to locate 3D objects.”
From Ten Years of Active Learning Techniques and Object Detection: A Systematic Review (2023) · cites this paper as [53]
Several works mentioned the inclusion of ensembles usage as part of active learning heuristics.
“Several works mentioned the inclusion of ensembles usage as part of AL heuristics [37,53,54].”
From Ten Years of Active Learning Techniques and Object Detection: A Systematic Review (2023) · cites this paper as [53]
Informativeness or uncertainty is calculated using Shannon Entropy and Mutual Information to analyse different points of view of epistemic uncertainty.
“Later, the informativeness or Uncertainty is calculated using thShannon Entropy and Mutual Information to analyse different points of view of the epistemic uncertainty, such as predictive uncertainty and confidence in the data, respectively.”
From Ten Years of Active Learning Techniques and Object Detection: A Systematic Review (2023) · cites this paper as [53]
MC-dropout and Deep Ensembles, when used to estimate uncertainty, show slightly better results compared to using a single softmax output.
“It is reported that MC-dropout and Deep Ensembles, when used to estimate Uncertainty, show slightly better results compared to using a single softmax output [53].”
From Ten Years of Active Learning Techniques and Object Detection: A Systematic Review (2023) · cites this paper as [53]
Cites
On calibration of modern neural networks
Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017) · arXiv · doi:10.48550/arxiv.1706.04599 · Cited by 3,435 · see citing works
Frustum PointNets for 3D Object Detection from RGB-D Data
Qi, C. R., Liu, W., Wu, C., Su, H., & Guibas, L. J. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00102 · Cited by 2,655 · see citing works
PIXOR: Real-time 3D Object Detection from Point Clouds
Yang, B., Luo, W., & Urtasun, R. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00798 · Cited by 1,249 · see citing works
The Power of Ensembles for Active Learning in Image Classification
Beluch, W. H., Genewein, T., Nurnberger, A., & Kohler, J. M. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00976 · Cited by 655 · see citing works
Vehicle Detection from 3D Lidar Using Fully Convolutional Network
Li, B., Zhang, T., & Xia, T. (2016) · Robotics: Science and Systems XII · doi:10.15607/rss.2016.xii.042 · Cited by 577 · see citing works
Deep Bayesian Active Learning with Image Data
Ghahramani, Z., Gal, Y., & Islam, R. (2017) · PMLR · doi:10.17863/cam.11070 · Cited by 392 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Feng, D., Haase-Schutz, C., Rosenbaum, L., Hertlein, H., Glaser, C., Timm, F., Wiesbeck, W., & Dietmayer, K. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2020.2972974 · Cited by 1,425 · see citing works
Self-citationA Survey of Deep Active Learning
Ren, P., Xiao, Y., Chang, X., Huang, P.-Y., Li, Z., Gupta, B. B., Chen, X., & Wang, X. (2022) · ACM Computing Surveys · doi:10.1145/3472291 · Cited by 1,169 · see citing works
Active Learning for Deep Object Detection via Probabilistic Modeling
Choi, J., Elezi, I., Lee, H.-J., Farabet, C., & Alvarez, J. M. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01010 · Cited by 135 · see citing works
“These methods reach good results but come with several limitations [11, 16].”
BackgroundMethodInfluentialDeep Active Learning for Computer Vision: Past and Future
Takezoe, R., Liu, X., Mao, S., Chen, M. T., Feng, Z., Zhang, S., & Wang, X. (2023) · APSIPA Transactions on Signal and Information Processing · doi:10.1561/116.00000057 · Cited by 26 · see citing works
Ten Years of Active Learning Techniques and Object Detection: A Systematic Review
Garcia, D., Carias, J., Adão, T., Jesus, R., Cunha, A., & Magalhães, L. G. (2023) · Applied Sciences · doi:10.3390/app131910667 · Cited by 17 · see citing works
- Authors
- Feng, D., Wei, X., Rosenbaum, L., Maki, A., & Dietmayer, K.
- Affiliations
- Universität Ulm (DE)
- School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology, Stockholm, 10044, Sweden
- Robert Bosch (Germany) (DE)
- KTH Royal Institute of Technology (SE)
- Venue
- 2019 IEEE Intelligent Vehicles Symposium (IV)
- Publisher
- IEEE
- Published
- 2019-06
- Pages
- 667-674
- ISSN
- 1931-0587 (ISSN-L)
- OpenAlex
- W2970624554
- Cited by
- 85
- References
- 6
- Citations by year
- 2026: 3 · 2025: 13 · 2024: 18 · 2023: 18 · 2022: 8 · 2021: 13 · 2020: 11 · 2019: 1
- FWCI
- 3.16
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Intelligent Vehicles Symposium (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Algorithms
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Keywords
- active learning (machine learning), autonomous driving, uncertainty estimation, query strategies, annotation cost reduction, point clouds, radar data
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
