3D LiDAR has become an indispensable sensor in autonomous driving vehicles. In LiDAR-based 3D point cloud semantic segmentation, most voxel-based 3D segmentors cannot efficiently capture large amounts of context information, resulting in limited receptive fields and limiting their performance. To address this problem, a sparse voxel-based attention network is introduced for 3D LiDAR point cloud semantic segmentation, termed SVASeg, which captures large amounts of context information between voxels through sparse voxel-based multi-head attention (SMHA). The traditional multi-head attention cannot directly be applied to the non-empty sparse voxels. To this end, a hash table is built according to the incrementation of voxel coordinates to lookup the non-empty neighboring voxels of each sparse voxel. Then, the sparse voxels are grouped into different groups, and each group corresponds to a local region. Afterwards, position embedding, multi-head attention and feature fusion are performed for each group to capture and aggregate the context information. Based on the SMHA module, the SVASeg can directly operate on the non-empty voxels, maintaining a comparable computational overhead to the convolutional method. Extensive experimental results on the SemanticKITTI and nuScenes datasets show the superiority of SVASeg.

It is extremely important and necessary for low computing power or portable devices to design more lightweight algorithms for image super-resolution (SR). Recently, most SR methods have achieved outstanding performance by sacrificing computational cost and memory storage, or vice versa. To address this problem, we introduce a lightweight U-shaped residual network (URNet) for fast and accurate image SR. Specifically, we propose a more effective feature distillation pyramid residual group (FDPRG) to extract features from low-resolution images. The FDPRG can effectively reuse the learned features with dense shortcuts and capture multi-scale information with a cascaded feature pyramid block. Based on the U-shaped structure, we utilize a step-by-step fusion strategy to improve the performance of feature fusion of different blocks. This strategy is different from the general SR methods which only use a single $Concat$ operation to fuse the features of all basic blocks. Moreover, a lightweight asymmetric residual non-local block is proposed to model the global context information and further improve the performance of SR. Finally, a high-frequency loss function is designed to alleviate smoothing image details caused by pixel-wise loss. Simultaneously, the proposed modules and high-frequency loss function can be easily plugged into multiple mature architectures to improve the performance of SR. Extensive experiments on multiple natural image datasets and remote sensing image datasets show the URNet achieves a better trade-off between image SR performance and model complexity against other state-of-the-art SR methods.

Traffic sign detection is a very important part of intelligent assisted driving system. However, with the interference of various target sizes, geometric distortion, occlusion and motion blur, fast and accurate detection on large-size car camera image is extremely hard. To achieve both high efficient and accurate detection, we present a traffic sign detection method within a coarse-to-fine framework, which sequentially detects the targets in grid-level and image-level. We demonstrate that focusing first is a more effective detection strategy for small targets in wide detection space. We propose a target grid prediction network, which is a fully convolutional network for binary classification, to realize rapid coarse localization of the target and effectively guide the clipping and scaling of the target area.With the flexible potential target region extracting strategy, the detecting space can be significantly reduced. At the same time, the correctly extracted local areas for the targets can further facilitate the accurate detection of the subsequent traffic sign detector. In the experiments, our method achieves impressive performance in terms of both efficiency and accuracy. On the challenging Tsinghua-Tencent 100K (TT100K) dataset, our method achieves 20.9 FPS detection speed for 1600 × 1600 images and the F1-score of the proposed method for all-scale targets is 91.55.

Current object detection algorithms perform inference on all samples at a fixed computational cost in the inference stage, which wastes computing resources and is not flexible. To solve this problem, a dynamic object detection algorithm based on a lightweight shared feature pyramid is proposed, which performs adaptive inference according to computing resources and the difficulty of samples, greatly improving the efficiency of inference. Specifically, a lightweight shared feature pyramid network and lightweight detection head is proposed to reduce the amount of computation and parameters in the feature fusion part and detection head of the dynamic object detection model. On the PASCAL VOC dataset, under the two conditions of “anytime prediction” and “budgeted batch object detection”, the performance, computation amount and parameter amount are better than the dynamic object detection models constructed by networks such as ResNet, DenseNet and MSDNet.

The Gaussian mixture model (GMM)-based methods have achieved great success in point set registration. However, they cannot be directly applied to image matching, because the features extracted from two images usually contain a large portion of outliers. In this letter, we propose a new method to extend the powerful GMM to the field of image feature points matching. The algorithm consists of two main steps. In the first step, points extracted from the images are mapped into a new subspace, in which feature similarity information is fused to get the new representation of the points. The second step performs an improved progressive process with the GMM to find correspondences satisfying the coherent constraint. In this way, finding correspondences among large outliers is feasible and the iteration converges faster. Experimental results on benchmark data sets show that the proposed method can find more correct matches with high accuracy.

Video segmentation is the foundation of many senior computer vision applications. For robotic vision applications, video object segmentation is facing more difficulties. Since the videos are usually collected from those complicated scenes and the camera perspective may be static for some time, object saliency and motion clues which are widely used in traditional video segmentation approaches could be unreliable. Besides, its online processing ability is of crucial importance. In this paper, we propose an online video object segmentation method, which is organized as a segmentation-by-detection framework. It first effectively segments out meaningful objects in short video frame batches with the help of object detectors, and then the segmentation of different batches can be associated with each other via bi-directional notebook based connecting strategy. Experiments conducted on public available datasets verify the above-mentioned property and show good performance of the proposed method.