OccTransformer:改进 BEVFormer 以实现仅基于相机的 3D 占据预测
OccTransformer: Improving BEVFormer for 3D camera-only occupancy prediction
AI总结:
本文提出 occTransformer,通过数据增强、强图像骨干、3D unet 头、多损失、模型集成及 StreamPETR 改进 BEVFormer,在 CVPR 2023 3D 占据预测赛道取得 49.23 miou。
摘要:
本技术报告介绍了我们为 CVPR 2023 自动驾驶挑战赛 3D 占据预测赛道提出的解决方案“occTransformer”。我们的方法建立在强基线 BEVFormer 之上,并通过若干简单而有效的技术提升其性能。首先,我们采用数据增强来增加训练数据的多样性,并提高模型的泛化能力。其次,我们使用强大的图像骨干网络从输入数据中提取信息量更丰富的特征。第三,我们加入一个 3D unet 头,以更好地捕捉场景的空间信息。第四,我们添加更多损失函数,以更好地优化模型。此外,我们使用与占据模型 BevDet 和 SurroundOcc 的集成方法来进一步提升性能。最重要的是,我们整合了 3D 检测模型 StreamPETR,以增强模型检测场景中物体的能力。通过使用这些方法,我们的解决方案在自动驾驶挑战赛的 3D 占据预测赛道上取得了 49.23 miou。
英文摘要:
This technical report presents our solution, "occTransformer" for the 3D occupancy prediction track in the autonomous driving challenge at CVPR 2023. Our method builds upon the strong baseline BEVFormer and improves its performance through several simple yet effective techniques. Firstly, we employed data augmentation to increase the diversity of the training data and improve the model's generalization ability. Secondly, we used a strong image backbone to extract more informative features from the input data. Thirdly, we incorporated a 3D unet head to better capture the spatial information of the scene. Fourthly, we added more loss functions to better optimize the model. Additionally, we used an ensemble approach with the occ model BevDet and SurroundOcc to further improve the performance. Most importantly, we integrated 3D detection model StreamPETR to enhance the model's ability to detect objects in the scene. Using these methods, our solution achieved 49.23 miou on the 3D occupancy prediction track in the autonomous driving challenge.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。