3-D Laser-Based Multiclass and Multiview Object Detection in Cluttered Indoor Scenes
- Submitting institution
-
The University of Essex
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 1149
- Type
- D - Journal article
- DOI
-
10.1109/TNNLS.2015.2496195
- Title of journal
- IEEE Transactions on Neural Networks and Learning Systems
- Article number
- -
- First page
- 177
- Volume
- 28
- Issue
- 1
- ISSN
- 2162-2388
- Open access status
- Out of scope for open access requirements
- Month of publication
- December
- Year of publication
- 2015
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
3
- Research group(s)
-
D - Robotics and Embedded Systems (RES)
- Citation count
- 8
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Published in IEEE-TNNLS, one of the top journals in the field, this paper presents a novel 3D object detection system using laser point clouds, which effectively deals with cluttered indoor scenes with imbalanced training data; an important research topic. Raw 3D point clouds are first transformed to 2D bearing angle images to reduce the computational cost. Then jointly trained multi-object detectors are deployed to perform multi-class and multi-view 3D object detection. The proposed system outperformed existing systems and made a good impact on service robots operated in a cluttered indoor environment. A further enhanced prototype has subsequently been created.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -