DriPE: A Dataset for Human Pose Estimation in Real-World Driving Settings

Guesdon, Romain; Crispim-Junior, Carlos; Tougne, Laure

doi:10.1109/iccvw54120.2021.00321

Cited by 13 publications

(1 citation statement)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This leads to a high accuracy, but also to a higher inference time. Guesdon et al [27] does further research with established state of the art top-down techniques for single-person driver pose estimation without an additional person detector. The head of our human pose estimation branch is based on the method of Cao et al [26] which includes several stages for predicting pairwise relationships (partial affinity fields) between keypoints, followed by a final stage for predicting the keypoints themselves.…”

Section: Related Workmentioning

confidence: 99%

Multitask Network for Joint Object Detection, Semantic Segmentation and Human Pose Estimation in Vehicle Occupancy Monitoring

Nikolas¹,

Mangat²,

Wasenmüller³

2022

Preprint

View full text Add to dashboard Cite

In order to ensure safe autonomous driving, precise information about the conditions in and around the vehicle must be available. Accordingly, the monitoring of occupants and objects inside the vehicle is crucial. In the state-of-the-art, single or multiple deep neural networks are used for either object recognition, semantic segmentation, or human pose estimation. In contrast, we propose our Multitask Detection, Segmentation and Pose Estimation Network (MDSP) -the first multitask network solving all these three tasks jointly in the area of occupancy monitoring. Due to the shared architecture, memory and computing costs can be saved while achieving higher accuracy. Furthermore, our architecture allows a flexible combination of the three mentioned tasks during a simple endto-end training. We perform comprehensive evaluations on the public datasets SVIRO and TiCaM in order to demonstrate the superior performance.

show abstract

Section: Related Workmentioning

confidence: 99%