Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | Proceedings - 35th IEEE Intelligent Vehicles Symposium, IV 2024 |
| Herausgeber (Verlag) | Institute of Electrical and Electronics Engineers Inc. |
| Seiten | 1236-1241 |
| Seitenumfang | 6 |
| ISBN (elektronisch) | 9798350348811 |
| ISBN (Print) | 979-8-3503-4882-8 |
| Publikationsstatus | Veröffentlicht - 2 Juni 2024 |
| Veranstaltung | 35th IEEE Intelligent Vehicles Symposium, IV 2024 - Jeju Island, Südkorea Dauer: 2 Juni 2024 → 5 Juni 2024 |
Publikationsreihe
| Name | IEEE Intelligent Vehicles Symposium, Proceedings |
|---|---|
| ISSN (Print) | 1931-0587 |
| ISSN (elektronisch) | 2642-7214 |
Abstract
Collective Perception has attracted significant attention in recent years due to its advantage for mitigating occlusion and expanding the field-of-view, thereby enhancing reliability, efficiency, and, most crucially, decision-making safety. However, developing collective perception models is highly resource demanding due to extensive requirements of processing input data for many agents, usually dozens of images and point clouds for a single frame. This not only slows down the model development process for collective perception but also impedes the utilization of larger models. In this paper, we propose an agent-based training framework that handles the deep learning modules and agent data separately to have a cleaner data flow structure. This framework not only provides an API for flexibly prototyping the data processing pipeline and defining the gradient calculation for each agent, but also provides the user interface for interactive training, testing and data visualization. Training experiment results of four collective object detection models on the prominent collective perception benchmark OPV2V show that the agent-based training can significantly reduce the GPU memory consumption and training time while retaining inference performance. The framework and model implementations are available at https://github.com/YuanYunshuang/CoSense3D
ASJC Scopus Sachgebiete
- Informatik (insg.)
- Angewandte Informatik
- Ingenieurwesen (insg.)
- Fahrzeugbau
- Mathematik (insg.)
- Modellierung und Simulation
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
Proceedings - 35th IEEE Intelligent Vehicles Symposium, IV 2024. Institute of Electrical and Electronics Engineers Inc., 2024. S. 1236-1241 (IEEE Intelligent Vehicles Symposium, Proceedings).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - CoSense3D
T2 - 35th IEEE Intelligent Vehicles Symposium, IV 2024
AU - Yuan, Yunshuang
AU - Sester, Monika
N1 - Publisher Copyright: © 2024 IEEE.
PY - 2024/6/2
Y1 - 2024/6/2
N2 - Collective Perception has attracted significant attention in recent years due to its advantage for mitigating occlusion and expanding the field-of-view, thereby enhancing reliability, efficiency, and, most crucially, decision-making safety. However, developing collective perception models is highly resource demanding due to extensive requirements of processing input data for many agents, usually dozens of images and point clouds for a single frame. This not only slows down the model development process for collective perception but also impedes the utilization of larger models. In this paper, we propose an agent-based training framework that handles the deep learning modules and agent data separately to have a cleaner data flow structure. This framework not only provides an API for flexibly prototyping the data processing pipeline and defining the gradient calculation for each agent, but also provides the user interface for interactive training, testing and data visualization. Training experiment results of four collective object detection models on the prominent collective perception benchmark OPV2V show that the agent-based training can significantly reduce the GPU memory consumption and training time while retaining inference performance. The framework and model implementations are available at https://github.com/YuanYunshuang/CoSense3D
AB - Collective Perception has attracted significant attention in recent years due to its advantage for mitigating occlusion and expanding the field-of-view, thereby enhancing reliability, efficiency, and, most crucially, decision-making safety. However, developing collective perception models is highly resource demanding due to extensive requirements of processing input data for many agents, usually dozens of images and point clouds for a single frame. This not only slows down the model development process for collective perception but also impedes the utilization of larger models. In this paper, we propose an agent-based training framework that handles the deep learning modules and agent data separately to have a cleaner data flow structure. This framework not only provides an API for flexibly prototyping the data processing pipeline and defining the gradient calculation for each agent, but also provides the user interface for interactive training, testing and data visualization. Training experiment results of four collective object detection models on the prominent collective perception benchmark OPV2V show that the agent-based training can significantly reduce the GPU memory consumption and training time while retaining inference performance. The framework and model implementations are available at https://github.com/YuanYunshuang/CoSense3D
KW - Collective Perception
KW - Efficient Training
KW - Object Detection
UR - http://www.scopus.com/inward/record.url?scp=85199788477&partnerID=8YFLogxK
U2 - 10.48550/arXiv.2404.18617
DO - 10.48550/arXiv.2404.18617
M3 - Conference contribution
SN - 979-8-3503-4882-8
T3 - IEEE Intelligent Vehicles Symposium, Proceedings
SP - 1236
EP - 1241
BT - Proceedings - 35th IEEE Intelligent Vehicles Symposium, IV 2024
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 2 June 2024 through 5 June 2024
ER -