Autonomous driving systems require extensive testing under a wide range of traffic conditions to ensure safety and reliability. Powered two-wheelers (PTWs), such as motorcycles and scooters, are particularly challenging due to their agile, nonlinear, and sometimes abrupt maneuvers. This paper proposes a conditional denoising diffusion probabilistic model to generate high-risk PTW pre-crash trajectories from limited in-depth crash data. By reversing a controlled noise process, the model synthesizes PTW trajectories that match the kinematic characteristics of real pre-crash cases while covering a wide range of risky interactions. Experiments on reconstructed crash trajectories from the China In-depth Mobility Safety Study Traffic Accident (CIMSS-TA) database show that the proposed method produces trajectories with higher physical plausibility, closer similarity to real-world crashes, and richer high-risk coverage than baseline models. When integrated into a simulation platform, the generated scenarios trigger more frequent and more diverse PTW conflict situations than the original crash set, indicating improved coverage of safety-critical conditions for autonomous vehicle testing.

Segmenting dental plaque from images of medical reagent staining provides valuable information for diagnosis and the determination of follow-up treatment plan. However, accurate dental plaque segmentation is a challenging task that requires identifying teeth and dental plaque subjected to semantic-blur regions (i.e., confused boundaries in border regions between teeth and dental plaque) and complex variations of instance shapes, which are not fully addressed by existing methods. Therefore, we propose a semantic decomposition network (SDNet) that introduces two single-task branches to separately address the segmentation of teeth and dental plaque and designs additional constraints to learn category-specific features for each branch, thus facilitating the semantic decomposition and improving the performance of dental plaque segmentation. Specifically, SDNet learns two separate segmentation branches for teeth and dental plaque in a divide-and-conquer manner to decouple the entangled relation between them. Each branch that specifies a category tends to yield accurate segmentation. To help these two branches better focus on category-specific features, two constraint modules are further proposed: 1) contrastive constraint module (CCM) to learn discriminative feature representations by maximizing the distance between different category representations, so as to reduce the negative impact of semantic-blur regions on feature extraction; 2) structural constraint module (SCM) to provide complete structural information for dental plaque of various shapes by the supervision of an boundary-aware geometric constraint. Besides, we construct a large-scale open-source Stained Dental Plaque Segmentation dataset (SDPSeg), which provides high-quality annotations for teeth and dental plaque. Experimental results on SDPSeg datasets show SDNet achieves state-of-the-art performance.