OAFuser: Towards Omni-Aperture Fusion for Light Field Semantic Segmentation

IEEE Transactions on Artificial Intelligence, 2024

Fei Teng · Jiaming Zhang · Kunyu Peng · Yaonan Wang · Rainer Stiefelhagen · Kailun Yang

Paper

🛠️ 👷 🚀

🔥 This repository is an integration of OAFuser and LFTrancy. 🔥

Update

• 2024.08.04 This repository for OAFuser is released.
• 2023.09.25 Codestuff is on processing.
• 2023.07.29 Init repository.
• 2023.07.31 Release the arXiv version.

TODO List

• Release the arXiv version.
• The code for OAFuser has been released.
• The integration of OAFuser and LFTracy will be released.
• Train and Eval strategy will be released.
• Checkpoints will be released.

Abstract

Light field cameras can provide rich angular and spatial information to enhance image semantic segmentation for scene understanding in the field of autonomous driving. However, the extensive angular information of light field cameras contains a large amount of redundant data, which is overwhelming for the limited hardware resource of intelligent vehicles. Besides, inappropriate compression leads to information corruption and data loss. To excavate representative information, we propose an Omni-Aperture Fusion model (OAFuser), which leverages dense context from the central view and discovers the angular information from sub-aperture images to generate a semantically-consistent result. To avoid feature loss during network propagation and simultaneously streamline the redundant information from the light field camera, we present a simple yet very effective Sub-Aperture Fusion Module (SAFM) to embed sub-aperture images into angular features without any additional memory cost. Furthermore, to address the mismatched spatial information across viewpoints, we present Center Angular Rectification Module (CARM) realized feature resorting and prevent feature occlusion caused by asymmetric information.

Method

(Overview)

🤝 Publication:

Please consider referencing this paper if you use the code or data from our work. Thanks a lot :)

@article{teng2024oafuser,
  title={OAFuser: Towards omni-aperture fusion for light field semantic segmentation of road scenes},
  author={Teng, Fei and Zhang, Jiaming and Peng, Kunyu and Wang, Yaonan and Stiefelhagen, Rainer and Yang, Kailun},
  journal={IEEE Transactions on Artificial Intelligence},
  year={2024}
}