Raw Image Segmentation (mode: Everything) RoI road sign, China ETC sign, overlapped cars pedestrain, indoor parking area image blurring, cars in shadow, railway potholes (drivable) holes (non-drivable)
Lift, Splat, Shoot: Encoding Images from Arbitrary Camera Rigs by Implicitly Unprojecting to 3D
We link 2D feature extraction and 3D object prediction via geometric back-projection with camera transfor-mation matrices.
论文名称:UniFormer: Unified Multi-view Fusion Transformer for Spatial-Temporal Representation in Bird’s-Eye-View
论文名称:Polar Parametrization for Vision-based Surround-View 3D Detection
论文名称:Occupancy Networks: Learning 3D Reconstruction in Function Space
论文名称:Exploring the Effects of Data Augmentation for Drivable Area Segmentation,是一篇Cityscapes数据集上构建freespace segmentation的技术报告。研究了语义分割模型的数据增强方法,提出了一个魔改 UNet + CBAM 模型,在 Cityscapes 验证,但没有说明是如何在 Cityscapes 构建 drivable area ground truth 的。
AssembleNet: Searching for Multi-Stream Neural Connectivity in Video Architectures (Google 作者AJ 发表于ICLR2020)