No Arabic abstract
In this paper, we present an Intersection-over-Union (IoU) guided two-stage 3D object detector with a voxel-to-point decoder. To preserve the necessary information from all raw points and maintain the high box recall in voxel based Region Proposal Network (RPN), we propose a residual voxel-to-point decoder to extract the point features in addition to the map-view features from the voxel based RPN. We use a 3D Region of Interest (RoI) alignment to crop and align the features with the proposal boxes for accurately perceiving the object position. The RoI-Aligned features are finally aggregated with the corner geometry embeddings that can provide the potentially missing corner information in the box refinement stage. We propose a simple and efficient method to align the estimated IoUs to the refined proposal boxes as a more relevant localization confidence. The comprehensive experiments on KITTI and Waymo Open Dataset demonstrate that our method achieves significant improvements with novel architectures against the existing methods. The code is available on Github URLfootnote{url{https://github.com/jialeli1/From-Voxel-to-Point}}.
Monocular 3D detection currently struggles with extremely lower detection rates compared to LiDAR-based methods. The poor accuracy is mainly caused by the absence of accurate location cues due to the ill-posed nature of monocular imagery. LiDAR point clouds, which provide precise spatial measurement, can offer beneficial information for the training of monocular methods. To make use of LiDAR point clouds, prior works project them to form depth map labels, subsequently training a dense depth estimator to extract explicit location features. This indirect and complicated way introduces intermediate products, i.e., depth map predictions, taking much computation costs as well as leading to suboptimal performances. In this paper, we propose LPCG (LiDAR point cloud guided monocular 3D object detection), which is a general framework for guiding the training of monocular 3D detectors with LiDAR point clouds. Specifically, we use LiDAR point clouds to generate pseudo labels, allowing monocular 3D detectors to benefit from easy-collected massive unlabeled data. LPCG works well under both supervised and unsupervised setups. Thanks to a general design, LPCG can be plugged into any monocular 3D detector, significantly boosting the performance. As a result, we take the first place on KITTI monocular 3D/BEV (birds-eye-view) detection benchmark with a considerable margin. The code will be made publicly available soon.
Due to the sparsity and irregularity of the 3D data, approaches that directly process points have become popular. Among all point-based models, Transformer-based models have achieved state-of-the-art performance by fully preserving point interrelation. However, most of them spend high percentage of total time on sparse data accessing (e.g., Farthest Point Sampling (FPS) and neighbor points query), which becomes the computation burden. Therefore, we present a novel 3D Transformer, called Point-Voxel Transformer (PVT) that leverages self-attention computation in points to gather global context features, while performing multi-head self-attention (MSA) computation in voxels to capture local information and reduce the irregular data access. Additionally, to further reduce the cost of MSA computation, we design a cyclic shifted boxing scheme which brings greater efficiency by limiting the MSA computation to non-overlapping local boxes while also preserving cross-box connection. Our method fully exploits the potentials of Transformer architecture, paving the road to efficient and accurate recognition results. Evaluated on classification and segmentation benchmarks, our PVT not only achieves strong accuracy but outperforms previous state-of-the-art Transformer-based models with 9x measured speedup on average. For 3D object detection task, we replace the primitives in Frustrum PointNet with PVT layer and achieve the improvement of 8.6%.
3D object detection is receiving increasing attention from both industry and academia thanks to its wide applications in various fields. In this paper, we propose the Point-Voxel Region based Convolution Neural Networks (PV-RCNNs) for accurate 3D detection from point clouds. First, we propose a novel 3D object detector, PV-RCNN-v1, which employs the voxel-to-keypoint scene encoding and keypoint-to-grid RoI feature abstraction two novel steps. These two steps deeply incorporate both 3D voxel CNN and PointNet-based set abstraction for learning discriminative point-cloud features. Second, we propose a more advanced framework, PV-RCNN-v2, for more efficient and accurate 3D detection. It consists of two major improvements, where the first one is the sectorized proposal-centric strategy for efficiently producing more representative and uniformly distributed keypoints, and the second one is the VectorPool aggregation to replace set abstraction for better aggregating local point-cloud features with much less resource consumption. With these two major modifications, our PV-RCNN-v2 runs more than twice as fast as the v1 version while still achieving better performance on the large-scale Waymo Open Dataset with 150m * 150m detection range. Extensive experiments demonstrate that our proposed PV-RCNNs significantly outperform previous state-of-the-art 3D detection methods on both the Waymo Open Dataset and the highly-competitive KITTI benchmark.
We present Point-Voxel CNN (PVCNN) for efficient, fast 3D deep learning. Previous work processes 3D data using either voxel-based or point-based NN models. However, both approaches are computationally inefficient. The computation cost and memory footprints of the voxel-based models grow cubically with the input resolution, making it memory-prohibitive to scale up the resolution. As for point-based networks, up to 80% of the time is wasted on structuring the sparse data which have rather poor memory locality, not on the actual feature extraction. In this paper, we propose PVCNN that represents the 3D input data in points to reduce the memory consumption, while performing the convolutions in voxels to reduce the irregular, sparse data access and improve the locality. Our PVCNN model is both memory and computation efficient. Evaluated on semantic and part segmentation datasets, it achieves much higher accuracy than the voxel-based baseline with 10x GPU memory reduction; it also outperforms the state-of-the-art point-based models with 7x measured speedup on average. Remarkably, the narrower version of PVCNN achieves 2x speedup over PointNet (an extremely efficient model) on part and scene segmentation benchmarks with much higher accuracy. We validate the general effectiveness of PVCNN on 3D object detection: by replacing the primitives in Frustrum PointNet with PVConv, it outperforms Frustrum PointNet++ by 2.4% mAP on average with 1.5x measured speedup and GPU memory reduction.
Mesh reconstruction from a 3D point cloud is an important topic in the fields of computer graphic, computer vision, and multimedia analysis. In this paper, we propose a voxel structure-based mesh reconstruction framework. It provides the intrinsic metric to improve the accuracy of local region detection. Based on the detected local regions, an initial reconstructed mesh can be obtained. With the mesh optimization in our framework, the initial reconstructed mesh is optimized into an isotropic one with the important geometric features such as external and internal edges. The experimental results indicate that our framework shows great advantages over peer ones in terms of mesh quality, geometric feature keeping, and processing speed.