In this paper, we propose a novel 3D object detector KDA3D, which achieves high-precision and robust classification, segmentation, and localization with the help of key-point densification and multi-attention guidance. The proposed end-to-end neural network architecture takes LIDAR point clouds as the main inputs that can be optionally complemented by RGB images. It consists of three parts: part-1 segments 3D foreground points and generates reliable proposals; part-2 (optional) enhances point cloud density and reconstructs the more compact full-point feature map; part-3 refines 3D bounding boxes and adds semantic segmentation as extra supervision. Our designed lightweight point-wise and channel-wise attention modules can adaptively strengthen the “skeleton” and “distinctiveness” point-features to help feature learning networks capture more representative or finer patterns. The proposed key-point densification component can generate pseudo-point clouds containing target information from monocular images through the distance preference strategy and K-means clustering so as to balance the density distribution and enrich sparse features. Extensive experiments on the KITTI and nuScenes 3D object detection benchmarks show that our KDA3D produces state-of-the-art results while running in near real-time with a low memory footprint.

Stereo matching networks based on deep learning are widely developed and can obtain excellent disparity estimation. We present a new end-to-end fast deep learning stereo matching network in this work that aims to determine the corresponding disparity from two stereo image pairs. We extract the characteristics of the low-resolution feature images using the stacked hourglass structure feature extractor and build a multi-level detailed cost volume. We also use the edge of the left image to guide disparity optimization and sub-sample with the low-resolution data, ensuring excellent accuracy and speed at the same time. Furthermore, we design a multi-cross attention model for binocular stereo matching to improve the matching accuracy and achieve end-to-end disparity regression effectively. We evaluate our network on Scene Flow, KITTI2012, and KITTI2015 datasets, and the experimental results show that the speed and accuracy of our method are excellent.