Bo Liu

3D LIDAR point clouds are extensively utilized in various domains, and data augmentation techniques for these point clouds can enhance network model convergence during training while also reducing the requisite data volume. Notably, PolarMix represents a seminal contribution to data enhancement in the realm of 3D LIDAR point Clouds Semantic Segmentation. It markedly augments the number of instances per class through swapping and rotate-paste mechanisms. Rotate-paste involves rotating and pasting selected class instances around the Z-axis multiple times. However, when capturing real-world scenarios using LiDAR point clouds, a pronounced class imbalance is observed, wherein certain classes dominate in sample numbers, while others are sparsely represented. Regrettably, PolarMix overlooks this class imbalance, leading to unequal treatment of all classes. To rectify this, we introduce the Class-Balanced PolarMix (CB-PolarMix), which operates in a cascading manner to diversify the training distribution and further optimize data augmentation outcomes. The cornerstone of CB-PolarMix lies in its adaptive reinforcement of foreground classes based on their distribution patterns. More specifically, our approach tweaks the pasting process for each class contingent upon its historical prediction accuracy. Experimental results from the SemanticPOSS and SemanticKitti datasets, utilizing the MinkowskiNet and SPVCNN models respectively, underscore the efficacy of the proposed CB-PolarMix.

Lane detection is a common task in computer vision that involves identifying the boundaries of lanes on a road from an image or a video. Improving the accuracy of lane detection is of great help to advanced driver assistance systems and autonomous driving that help cars to identify and keep in the correct lane. Current high-accuracy models of lane detection are mainly based on artificial neural networks. Among them, CLRNet is the latest famous model, which attains high lane detection accuracy. However, in some scenarios, CLRNet attains lower lane detection accuracy, and we revealed that this is caused by insufficient global dependence information. In this study, we enhanced CLRNet and proposed a new model called NonLocal CLRNet (NLNet). NonLocal is an algorithmic mechanism that captures long-range dependence. NLNet employs NonLocal to acquire more long-range dependence information or global information and then applies the acquired information to a Feature Pyramid Network (FPN) in CLRNet for improving lane detection accuracy. Using the CULane dataset, we trained NLNet. The experimental results showed that NLNet outperformed state-of-the-art models in terms of accuracy in most scenarios, particularly in the no-line scenario and night scenario. This study is very helpful for developing more accurate lane detection models.