Yanxin Ma

Yifan Zhang, Qingyong Hu, Guoquan Xu et al.

We study the problem of efficient object detection of 3D LiDAR point clouds. To reduce the memory and computational cost, existing point-based pipelines usually adopt task-agnostic random sampling or farthest point sampling to progressively downsample input point clouds, despite the fact that not all points are equally important to the task of object detection. In particular, the foreground points are inherently more important than background points for object detectors. Motivated by this, we propose a highly-efficient single-stage point-based 3D detector in this paper, termed IA-SSD. The key of our approach is to exploit two learnable, task-oriented, instance-aware downsampling strategies to hierarchically select the foreground points belonging to objects of interest. Additionally, we also introduce a contextual centroid perception module to further estimate precise instance centers. Finally, we build our IA-SSD following the encoder-only architecture for efficiency. Extensive experiments conducted on several large-scale detection benchmarks demonstrate the competitive performance of our IA-SSD. Thanks to the low memory footprint and a high degree of parallelism, it achieves a superior speed of 80+ frames-per-second on the KITTI dataset with a single RTX2080Ti GPU. The code is available at \url{https://github.com/yifanzhang713/IA-SSD}.

Guoquan Xu, Hezhi Cao, Yifan Zhang et al.

Attention mechanism plays a more and more important role in point cloud analysis and channel attention is one of the hotspots. With so much channel information, it is difficult for neural networks to screen useful channel information. Thus, an adaptive channel encoding mechanism is proposed to capture channel relationships in this paper. It improves the quality of the representation generated by the network by explicitly encoding the interdependence between the channels of its features. Specifically, a channel-wise convolution (Channel-Conv) is proposed to adaptively learn the relationship between coordinates and features, so as to encode the channel. Different from the popular attention weight schemes, the Channel-Conv proposed in this paper realizes adaptability in convolution operation, rather than simply assigning different weights for channels. Extensive experiments on existing benchmarks verify our method achieves the state of the arts.

Guoquan Xu, Hezhi Cao, Yifan Zhang et al.

Transformer plays an increasingly important role in various computer vision areas and remarkable achievements have also been made in point cloud analysis. Since they mainly focus on point-wise transformer, an adaptive channel encoding transformer is proposed in this paper. Specifically, a channel convolution called Transformer-Conv is designed to encode the channel. It can encode feature channels by capturing the potential relationship between coordinates and features. Compared with simply assigning attention weight to each channel, our method aims to encode the channel adaptively. In addition, our network adopts the neighborhood search method of low-level and high-level dual semantic receptive fields to improve the performance. Extensive experiments show that our method is superior to state-of-the-art point cloud classification and segmentation methods on three benchmark datasets.

Guoquan Xu, Hezhi Cao, Yifan Zhang et al.

Recently, deep neural networks have made remarkable achievements in 3D point cloud classification. However, existing classification methods are mainly implemented on idealized point clouds and suffer heavy degradation of per-formance on non-idealized scenarios. To handle this prob-lem, a feature representation learning method, named Dual-Neighborhood Deep Fusion Network (DNDFN), is proposed to serve as an improved point cloud encoder for the task of non-idealized point cloud classification. DNDFN utilizes a trainable neighborhood learning method called TN-Learning to capture the global key neighborhood. Then, the global neighborhood is fused with the local neighbor-hood to help the network achieve more powerful reasoning ability. Besides, an Information Transfer Convolution (IT-Conv) is proposed for DNDFN to learn the edge infor-mation between point-pairs and benefits the feature transfer procedure. The transmission of information in IT-Conv is similar to the propagation of information in the graph which makes DNDFN closer to the human reasoning mode. Extensive experiments on existing benchmarks especially non-idealized datasets verify the effectiveness of DNDFN and DNDFN achieves the state of the arts.