It is difficult for lightweight neural networks to produce accurate 6DoF pose estimation effects due to their accuracy being affected by scale changes. To solve this problem, we propose a method with good performance and robustness based on previous research. The enhanced PVNet-based method uses depth-wise convolution to build a lightweight network. In addition, coordinate attention and atrous spatial pyramid pooling are used to ensure accuracy and robustness. This method effectively reduces the network size and computational complexity and is a lightweight 6DoF pose estimation method based on monocular RGB images. Experiments on public datasets and self-built datasets show that the average ADD(-S) estimation accuracy and 2D projection index of the improved method are improved. For datasets with large changes in object scale, the estimation accuracy of the average ADD(-S) is greatly improved.

The visual analysis of trajectory topics is helpful for mining potential trajectory patterns, but the traditional visual analysis method ignores the evolution of the temporal coherence of the topic. In this paper, a novel visual analysis method for dynamic topic analysis of traffic trajectory is proposed, which is used to explore and analyze the traffic trajectory topic and evolution. Firstly, the spatial information is integrated into trajectory words, calculating the dynamic trajectory topic model based on dynamic analysis modeling and, consequently, correlating the evolution of the trajectory topic between adjacent time slices. Secondly, in the trajectory topic, a representative trajectory sequence is generated to overcome the problem of the trajectory topic model not considering the word order, based on the improved Markov Decision Process. Subsequently, a set of meaningful visual codes is designed to analyze the trajectory topic and its evolution through the parallel window visual model from a spatial-temporal perspective. Finally, a case evaluation shows that the proposed method is effective in analyzing potential trajectory movement patterns.

Analyzing the evolution of trajectory topics is fundamental to understanding urban mobility and human activity. Existing methods, however, often struggle to capture complex spatio-temporal semantics and are constrained by fixed time windows, which limits multi-scale temporal analysis. This paper presents a novel method to model the dynamic topics of trajectories to address these limitations. The proposed method combines a domain-specific trajectory embedding strategy, a flexible dynamic topic modeling pipeline, and an interactive visualization system to address these limitations. Firstly, the method introduces a novel embedding method that uses a retrained RoBERTa model with a word-level tokenizer on Morton-coded trajectories to effectively learn spatial context and sequential patterns. Secondly, a BERTopic-based approach is employed for topic modeling, featuring an adjustable time window that allows for flexible analysis of topic dynamics across different temporal scales without model retraining. Furthermore, an interactive visualization system with coordinated spatio-temporal views translates abstract model outputs into an intuitive format, enabling direct exploration of evolving trajectory topics. Experiments on a large-scale taxi trajectory dataset demonstrate the proposed method’s effectiveness in identifying coherent and meaningful patterns of dynamic trajectory topics.

Online dynamic network visualization is imperative for real-time network monitoring and analysis applications. It presents a significant research challenge for maintaining both layout stability and quality amid unpredictable temporal changes. While node-link diagrams are extensively utilized in online dynamic network visualization, previous node-link-diagram-based research primarily focused on stabilizing the layout by defining constraints on local node movement. However, these constraints often neglect the structural influence and its corresponding global impact, which may lead to that the representations of the network structure change significantly over time and a decrease in layout quality. To address this problem, we introduce the Structure-based Influence Propagation and Aging (SIPA) algorithm, a novel approach to preserve the stability of relative node positions and shapes of interconnected nodes (referred to as structures) between adjacent time steps. These stable structures serve as visual cues for users tracking the evolution of the network, thereby enhancing the overall layout stability. Additionally, we enhance dynamic network analysis by a highly interactive visualization system, enriching the layout result with multiple coordinated views of temporal trends, network features, animated graph diaries and snapshots. Our approach empowers users to interactively track and compare network evolution within a long-term temporal context and across multiple aspects. We demonstrate the effectiveness and performance of our approach through in-lab user studies and comparative experiments with three baseline dynamic network layout methods.