FVNet: 3D Front-View Proposal Generation for Real-Time Object Detection from Point Clouds

Segmenting or detecting objects in sparse Lidar point clouds are two important tasks in autonomous driving to allow a vehicle to act safely in its 3D environment. The best performing methods in 3D semantic segmentation or object detection rely on a large amount of annotated data. Yet annotating 3D Lidar data for these tasks is tedious and costly. In this context, we propose a self-supervised pretraining method for 3D perception models that is tailored to autonomous driving data. Specifically, we leverage the availability of synchronized and calibrated image and Lidar sensors in autonomous driving setups for distilling self-supervised pre-trained image representations into 3D models. Hence, our method does not require any point cloud nor image annotations. The keyingredient of our method is the use of superpixels which are used to pool 3D point features and 2D pixel features in visually similar regions. We then train a 3D network on the self-supervised task of matching these pooled point features with the corresponding pooled image pixel features. The advantages of contrasting regions obtained by superpixels are that: (1) grouping together pixels and points of visually coherent regions leads to a more meaningful contrastive task that produces features well adapted to 3D semantic segmentation and 3D object detection; (2) all the different regions have the same weight in the contrastive loss regardless of the number of 3D points sampled in these regions; (3) it mitigates the noise produced by incorrect matching of points and pixels due to occlusions between the different sensors. Extensive experiments on autonomous driving datasets demonstrate the ability of our image-to-Lidar distillation strategy to produce 3D representations that transfer well on semantic segmentation and object detection tasks.

Accurate object detection is a fundamental requirement for autonomous systems to operate in dynamic urban environments. Considering the intricacy of the environment and the occlusion of objects, various point cloud based three-dimensional (3D) object detection methods have been proposed, such as point-based or voxel-based methods. In this paper, a feature selection mechanism is proposed in a voxel-based method to generate bird’s eye view (BEV) feature maps from the original point cloud. For the 3D Backbone, the U-shaped structure and single scale feature selection module are combined. After combining high-level semantics and low-level fine-grained features, the optimized features after the BEV feature maps are applied to region of interest (RoI) refinement, so that the voxel features can better serve the subsequent 3D object detection. The experimental results on KITTI dataset show higher 3D object detection accuracy compared to the state-of-the-art 3D detection methods, which reflects the effectiveness of the proposed architecture.

Dense frustum-aware fusion for 3D object detection in perception systems

Currently, there are many kinds of voxel-based multisensor 3D object detectors, while point-based multisensor 3D object detectors have not been fully studied. In this paper, we propose a new 3D two-stage object detection method based on point cloud and image fusion to improve the detection accuracy. To address the problem of insufficient semantic information of point cloud, we perform multiscale deep fusion of LiDAR point and camera image in a point-wise manner to enhance point features. Due to the imbalance of LiDAR points, the object point cloud in the long-distance area is sparse. We design a point cloud completion module to predict the spatial shape of objects in the candidate boxes and extract the structural information to improve the feature representation ability to further refine the boxes. The framework is evaluated on widely used KITTI and SUN-RGBD dataset. Experimental results show that our method outperforms all state-of-the-art point-based 3D object detection methods and has comparable performance to voxel-based methods as well.

Voxel is one of the common structural representation of 3D point cloud. Due to the sparsity of point cloud generated by light detection and ranging (LiDAR), there is the extreme imbalance in the foreground and background voxels. It decreases the accuracy of 3D object detection, has the negative effect on intelligent driving safety. To overcome this problem, we present a saliency prediction based 3D object detector SP-Det in this article. Although foreground voxels have the sufficient feature of object, it is difficult to localize the foreground region from voxel space with the larger background region. We design an auxiliary learning task, saliency prediction (SP). It benefits 3D detector in identifying the foreground region. SP task uses label diffusion to alleviate the label imbalance. It reduces the learning difficulty of saliency in voxel and bird's eye view (BEV) spaces. After that, to strengthen feature interaction from the sparse foreground region, we design saliency fusion (SF) module to fuse the learning result in SP task. It utilizes voxel and BEV saliency maps as progressive attention to resist the redundant feature from background region. To aggregate more foreground feature inside 3D and BEV region of interest (RoI), we design hybrid grid maps based RoI pooling (Hybrid-RoI pooling). Experiments are conducted in STF dataset. The adverse weather enlarges the sparsity of LiDAR point cloud, increasing the difficulty of object detection. SP-Det identifies and leverages foreground region, and achieves the performance better than the current methods. Hence, we believe that SP-Det benefits to LiDAR based 3D scene understanding in the adverse weather.

The 3D object detection is becoming indispensable for environmental perception in autonomous driving. Light detection and ranging (LiDAR) point clouds often fail to distinguish objects with similar structures and are quite sparse for distant or small objects, thereby introducing false and missed detections. To address these issues, LiDAR is often fused with cameras due to the rich textural information provided by images. However, current fusion methods suffer the inefficient data representation and inaccurate alignment of heterogeneous features, leading to poor precision and low efficiency. To this end, we propose a plug-and-play module termed range-image fusion (RI-Fusion) to achieve an effective fusion of LiDAR and camera data, designed to be easily accessible by existing mainstream LiDAR-based algorithms. In this process, we design an image and point cloud alignment method by converting a point cloud into a compact range-view representation through a spherical coordinate transformation. The range image is then integrated with a corresponding camera image utilizing an attention mechanism. The original range image is then concatenated with fusion features to retain point cloud information, and the results are projected onto a spatial point cloud. Finally, the feature-enhanced point cloud can be input into a LiDAR-based 3D object detector. The results of validation experiments involving the KITTI 3D object detection benchmark showed that our proposed fusion method significantly enhanced multiple mainstream LiDAR-based 3D object detectors, PointPillars, SECOND, and Part <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\text{A}{^{2}}$ </tex-math></inline-formula> , improving the 3D mAP (mean Average Precision) by 3.61%, 2.98%, and 1.27%, respectively, particularly for small objects such as pedestrians and cyclists.

In recent years, 3D object detection has attracted considerable interest from researchers due to its crucial importance in scene understanding tasks for mobile robots. The task of 3D object detection is to determine the position, orientation, scale, and category of objects in a scene, with multimodal data inputs. These inputs consist of two distinct types: images and point clouds. Images provide colour information about the scene, while point clouds deliver geometric information. The fusion of these two diverse data types can enhance the detection of 3D objects by allowing them to complement each other. In this paper, a review of real-time multimodal 3D object detection from the perspective of multimodal sensor fusion is presented. It is important to note that previous reviews of multimodal 3D object detection have been limited to autonomous driving scenarios. In contrast, this paper reviews multimodal 3D object detection approaches in both indoor and outdoor scenarios. In addition, the popular datasets and assessment metrics employed for 3D object detection are described, as well as the issues and future prospects of 3D object detection.

Light detection and ranging (LiDAR) is widely used in the automotive industry as it can provide point cloud information containing precise distances. Three-dimensional (3D) object detection based on LiDAR point clouds is significant for environment perception tasks. However, feature learning for point clouds remains challenging. This paper proposes a two-stage voxel-based LiDAR 3D object detector, referred to as density-aware and neighborhood attention (DenNet), that focuses on the neighborhood information of objects. DenNet mainly integrates two modules: voxel density-aware (VDA) and neighborhood attention (NA). VDA introduces density information of the point cloud. Here, point cloud density information was added as voxel features in the voxel-based framework to alleviate the information loss during voxelization. Additionally, to extract neighbor information, the characteristics of 3D objects were analyzed for traffic scenes. The NA mechanism was adopted, which localizes the receptive field for each query to its nearest neighboring points. DenNet yielded competitive results, as compared with state-of-the-art methods, for the KITTI and One Million Scenes (ONCE) datasets; notably, it afforded an improvement of 3.96% relative to the baseline mean average precision on the more challenging ONCE dataset.

Background: 3D object detection based on point clouds in road scenes has attracted much attention recently. The voxel-based methods voxelize the scene to regular grids, which can be processed with the advanced feature learning frameworks based on convolutional layers for semantic feature learning. The point-based methods can extract the geometric feature of the point due to the coordinate reservations. The combination of the two is effective for 3D object detection. However, the current methods use a voxel-based detection head with anchors for classification and localization. Although the preset anchors cover the entire scene, it is not suitable for detection tasks with larger scenes and multiple categories of objects, due to the limitation of the voxel size. Additionally, the misalignment between the predicted confidence and proposals in the Regions of the Interest (ROI) selection bring obstacles to 3D object detection. Methods: We investigate the combination of voxel-based methods and point-based methods for 3D object detection. Additionally, a voxel-to-point module that captures semantic and geometric features is proposed in the paper. The voxel-to-point module is conducive to the detection of small-size objects and avoids the presets of anchors in the inference stage. Moreover, a confidence adjustment module with the center-boundary-aware confidence attention is proposed to solve the misalignment between the predicted confidence and proposals in the regions of the interest selection. Results: The proposed method has achieved state-of-the-art results for 3D object detection in the Karlsruhe Institute of Technology and Toyota Technological Institute (KITTI) object detection dataset. Actually, as of September 19, 2021, our method ranked 1st in the 3D and Bird Eyes View (BEV) detection of cyclists tagged with difficulty level ‘easy’, and ranked 2nd in the 3D detection of cyclists tagged with ‘moderate’. Conclusions: We propose an end-to-end two-stage 3D object detector with voxel-to-point module and confidence adjustment module.

3D object detection is a core component of automated driving systems. State-of-the-art methods fuse RGB imagery and LiDAR point cloud data frame-by-frame for 3D bounding box regression. However, frame-by-frame 3D object detection suffers from noise, field-of-view obstruction, and sparsity. We propose a novel Temporal Fusion Module (TFM) to use information from previous time-steps to mitigate these problems. First, a state-of-the-art frustum network extracts point cloud features from raw RGB and LiDAR point cloud data frame-by-frame. Then, our TFM module fuses these features with a recurrent neural network. As a result, 3D object detection becomes robust against single frame failures and transient occlusions. Experiments on the KITTI object tracking dataset show the efficiency of the proposed TFM, where we obtain 6%, 4%, and 6% improvements on Car, Pedestrian, and Cyclist classes, respectively, compared to frame-by-frame baselines. Furthermore, ablation studies reinforce that the subject of improvement is temporal fusion and show the effects of different placements of TFM in the object detection pipeline. Our code is open-source and available at https://github.com/emecercelik/Temp-Frustum-Net.git.

Robust environmental perception systems are key to implementing advanced driver assistance features in modern vehicles. To acquire knowledge of the environment, vehicles employ sensors like LiDAR, Camera and Radar for object detection and tracking. For autonomous driving applications (SAE J3016 Level 3 and above), LiDAR and camera-based systems are most common. Camera-based systems have difficulty in operating under low-visibility conditions while LiDAR-based systems do not perform well in detecting objects at farther distances due to sparse point clouds. This work presents an approach to improving the shortcomings of the LiDAR-based detection systems by upsampling sparse LiDAR point clouds and then fusing the camera data to it to improve the quality of the point clouds generated that can be then used for object detection and identification. The proposed technique has been validated on the KITTI vision benchmark dataset. Results indicate that the proposed system is able to outperform standalone LiDAR-based 3D Object detection.

In this paper, we propose SparseDet for end-to-end 3D object detection from point cloud. Existing works on 3D object detection rely on dense object candidates over all locations in a 3D or 2D grid following the mainstream methods for object detection in 2D images. However, this dense paradigm requires expertise in data to fulfill the gap between label and detection. As a new detection paradigm, SparseDet maintains a fixed set of learnable proposals to represent latent candidates and directly perform classification and localization for 3D objects through stacked transformers. It demonstrates that effective 3D object detection can be achieved with none of post-processing such as redundant removal and non-maximum suppression. With a properly designed network, SparseDet achieves highly competitive detection accuracy while running with a more efficient speed of 34.5 FPS. We believe this end-to-end paradigm of SparseDet will inspire new thinking on the sparsity of 3D object detection.

We propose octree-based transformers, named OctFormer, for 3D point cloud learning. OctFormer can not only serve as a general and effective backbone for 3D point cloud segmentation and object detection but also have linear complexity and is scalable for large-scale point clouds. The key challenge in applying transformers to point clouds is reducing the quadratic, thus overwhelming, computation complexity of attentions. To combat this issue, several works divide point clouds into non-overlapping windows and constrain attentions in each local window. However, the point number in each window varies greatly, impeding the efficient execution on GPU. Observing that attentions are robust to the shapes of local windows, we propose a novel octree attention, which leverages sorted shuffled keys of octrees to partition point clouds into local windows containing a fixed number of points while permitting shapes of windows to change freely. And we also introduce dilated octree attention to expand the receptive field further. Our octree attention can be implemented in 10 lines of code with open-sourced libraries and runs 17 times faster than other point cloud attentions when the point number exceeds 200 k. Built upon the octree attention, OctFormer can be easily scaled up and achieves state-of-the-art performances on a series of 3D semantic segmentation and 3D object detection benchmarks, surpassing previous sparse-voxel-based CNNs and point cloud transformers in terms of both efficiency and effectiveness. Notably, on the challenging ScanNet200 dataset, OctFormer outperforms sparse-voxel-based CNNs by 7.3 in mIoU. Our code and trained models are available at https://wang-ps.github.io/octformer.

Safe self-driving vehicles require precise 3D Object Identification. LiDAR sensors are key in accomplishing such a task, as LiDARs produce high-definition point clouds. Such point clouds are then processed by 3D Object Detection models to finally detect objects.Most Object Detection models require massive amounts of data to be trained. Gathering and processing this data is an expensive and time-consuming task, which is why the information taken from each sample must be fully harnessed. Such can be done through Data Augmentation. Data Augmentation contributes significantly for improving performance, being at least as relevant as the advances in the Object Detection models themselves.A few studies have been reported regarding the effectiveness of Data Augmentation. However, the role played by original and augmented samples has been neglected. This work reports the first-ever detailed quantification of the impact that the inclusion of original and augmented samples in a dataset has in 3D Object Detection in the context of autonomous driving. The obtained results show that although a good augmentation strategy is crucial to the model’s performance, it is only as good as the quality of the original samples allows it to be.

3D object detection is critical for autonomous driving, enabling precise obstacle localization to support navigation and decision-making. However, detecting small objects remains challenging due to sparse point clouds and suboptimal multimodal fusion. To address these challenges, this paper proposes the sparse fuse dense with interlaced-attention and fusion network (SFD-IAFNet), a novel multimodal fusion-based 3D object detection framework. First, building on the SFD network to resolve the coarse feature extraction of pseudo point clouds, we design the multi-scale feature enhancement module VoxelFPN_CMLP module, voxel pyramid structure and compound multilayer perceptron are introduced, which can effectively alleviate the fuzzy problem of pseudo-point cloud in feature expression of long-distance targets, which deeply explores multi-level features of pseudo point clouds, significantly enhancing the feature representation of small objects. Second, for insufficient fusion and multimodal semantic alignment between raw and pseudo-point cloud features, we introduce the Interlaced-Attention mechanism, a mechanism that takes bimodal features as input and realizes deep fusion between pseudo-point cloud and raw point cloud through two-layer cross-decoding, which significantly improves fusion accuracy and context semantic modeling ability. Additionally, a dynamic loss weighting strategy is introduced to prioritize hard samples during training. Experimental results on the KITTI dataset demonstrate that our method achieves 95.84% (easy), 88.90% (mod), and 87.67% (hard) in 3D detection accuracy, outperforming the baseline SFD model by 0.36%, 1.43%, and 2.29%, respectively. Ablation studies validate the efficacy of our modules, particularly in complex scenarios with occlusions or distant small objects.