3D Object Detection

YOLOv2, or YOLO9000, is a single-stage real-time object detection model. It improves upon YOLOv1 in several ways, including the use of Darknet-19 as a backbone, batch normalization, use of a high-resolution classifier, and the use of anchor boxes to predict bounding boxes, and more.

Voxel RoI Pooling is a RoI feature extractor extracts RoI features directly from voxel features for further refinement. It starts by dividing a region proposal into $G \times G \times G$ regular sub-voxels. The center point is taken as the grid point of the corresponding sub-voxel. Since $3 D$ feature volumes are extremely sparse (non-empty voxels account for $<3 \%$ spaces), we cannot directly utilize max pooling over features of each sub-voxel. Instead, features are integrated from neighboring voxels into the grid points for feature extraction. Specifically, given a grid point $g\_{i}$, we first exploit voxel query to group a set of neighboring voxels $\Gamma\_{i}=\left\(\mathbf{v}\_{i}^{1}, \mathbf{v}\_{i}^{2}, \cdots, \mathbf{v}\_{i}^{K}\right\) .$ Then, we aggregate the neighboring voxel features with a PointNet module $\mathrm{a}$ as: $$ \mathbf{\eta}\_{i}=\max _{k=1,2, \cdots, K}\left\(\Psi\left(\left[\mathbf{v}\_{i}^{k}-\mathbf{g}\_{i} ; \mathbf{\phi}\_{i}^{k}\right]\right)\right\) $$ where $\mathbf{v}\_{i}-\mathbf{g}\_{i}$ represents the relative coordinates, $\mathbf{\phi}\_{i}^{k}$ is the voxel feature of $\mathbf{v}\_{i}^{k}$, and $\Psi(\cdot)$ indicates an MLP. The max pooling operation $\max (\cdot)$ is performed along the channels to obtain the aggregated feature vector $\eta_{i} .$ Particularly, Voxel RoI pooling is exploited to extract voxel features from the 3D feature volumes out of the last two stages in the $3 \mathrm{D}$ backbone network. And for each stage, two Manhattan distance thresholds are set to group voxels with multiple scales. Then, we concatenate the aggregated features pooled from different stages and scales to obtain the RoI features.

MotionNet is a system for joint perception and motion prediction based on a bird's eye view (BEV) map, which encodes the object category and motion information from 3D point clouds in each grid cell. MotionNet takes a sequence of LiDAR sweeps as input and outputs the bird's eye view (BEV) map. The backbone of MotionNet is a spatio-temporal pyramid network, which extracts deep spatial and temporal features in a hierarchical fashion. To enforce the smoothness of predictions over both space and time, the training of MotionNet is further regularized with novel spatial and temporal consistency losses.

Disp R-CNN is a 3D object detection system for stereo images. It utilizes an instance disparity estimation network (iDispNet) that predicts disparity only for pixels on objects of interest and learns a category-specific shape prior for more accurate disparity estimation. To address the challenge from scarcity of disparity annotation in training, a statistical shape model is used to generate dense disparity pseudo-ground-truth without the need of LiDAR point clouds.

Point-GNN is a graph neural network for detecting objects from a LiDAR point cloud. It predicts the category and shape of the object that each vertex in the graph belongs to. In Point-GNN, there is an auto-registration mechanism to reduce translation variance, as well as a box merging and scoring operation to combine detections from multiple vertices accurately.

VoTr is a Transformer-based 3D backbone for 3D object detection from point clouds. It contains a series of sparse and submanifold voxel modules. Submanifold voxel modules perform multi-head self-attention strictly on the non-empty voxels, while sparse voxel modules can extract voxel features at empty locations. Long-range relationships between voxels are captured via self-attention. Given the fact that non-empty voxels are naturally sparse but numerous, directly applying standard Transformer on voxels is non-trivial. To this end, VoTr uses a sparse voxel module and a submanifold voxel module, which can operate on the empty and non-empty voxel positions effectively. To further enlarge the attention range while maintaining comparable computational overhead to the convolutional counterparts, two attention mechanisms are used for multi-head attention in those two modules: Local Attention and Dilated Attention. Furthermore Fast Voxel Query is used to accelerate the querying process in multi-head attention.