TY - GEN
T1 - DCSN
T2 - 47th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2022
AU - Huang, Bor Sheng
AU - Hsu, Chih Chung
AU - Liao, Wo Ting
AU - Kao, Han Yi
AU - Wang, Xian Yun
N1 - Funding Information:
This study was supported in part by the Ministry of Science and Technology, Taiwan, under Grants MOST 109-2218-E-006-032, 107-2218-E-020-002-MY3, and 109-2634-F-007-013. We thank to National Center for High-performance Computing (NCHC) of National Applied Research Laboratories (NARLabs) in Taiwan for providing computational and storage resources.
Publisher Copyright:
© 2022 IEEE
PY - 2022
Y1 - 2022
N2 - This paper presents a novel semantic segmentation network for outdoor and unstructured scenarios for autonomous driving based on deformable convolution and geometric distortion pipelines. The semantic segmentation tasks for autonomous driving are generally designed for the urban scene, city-view, and highly structured scenarios, such as the CityScapes dataset, KITTI, and BDD, while rare study focuses on outskirts scenarios. Therefore, the performance of existing semantic segmentation networks on such datasets might be unreliable. To conquer this issue, a novel densely connected residual block (DCRB) with the deformable convolution is proposed to form our backbone for capturing the non-rigid feature representation. In this way, the gradient flow of our DCRB could be better back-propagated from the segmentation head, resulting in a stable training process. Second, geometric distortion augmentation is introduced in the data augmentation pipeline, simulating the possible deformation situations in real-world outdoor scenarios. The experiments are conducted that the proposed semantic segmentation network significantly outperforms the state-of-the-art methods for both Cityscapes and Outdoor scenarios.
AB - This paper presents a novel semantic segmentation network for outdoor and unstructured scenarios for autonomous driving based on deformable convolution and geometric distortion pipelines. The semantic segmentation tasks for autonomous driving are generally designed for the urban scene, city-view, and highly structured scenarios, such as the CityScapes dataset, KITTI, and BDD, while rare study focuses on outskirts scenarios. Therefore, the performance of existing semantic segmentation networks on such datasets might be unreliable. To conquer this issue, a novel densely connected residual block (DCRB) with the deformable convolution is proposed to form our backbone for capturing the non-rigid feature representation. In this way, the gradient flow of our DCRB could be better back-propagated from the segmentation head, resulting in a stable training process. Second, geometric distortion augmentation is introduced in the data augmentation pipeline, simulating the possible deformation situations in real-world outdoor scenarios. The experiments are conducted that the proposed semantic segmentation network significantly outperforms the state-of-the-art methods for both Cityscapes and Outdoor scenarios.
UR - http://www.scopus.com/inward/record.url?scp=85131253117&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85131253117&partnerID=8YFLogxK
U2 - 10.1109/ICASSP43922.2022.9747586
DO - 10.1109/ICASSP43922.2022.9747586
M3 - Conference contribution
AN - SCOPUS:85131253117
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 4668
EP - 4672
BT - 2022 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2022 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 23 May 2022 through 27 May 2022
ER -