Multi-view video reconstruction plays a vital role in computer vision, enabling applications in film production, virtual reality, and motion analysis. While recent advances such as 3D Gaussian Splatting have demonstrated impressive capabilities in dynamic scene reconstruction, they typically rely on the assumption that input video streams are temporally synchronized. However, in real-world scenarios, this assumption often fails due to factors like camera trigger delays, frame rate discrepancies, or independent recording setups, leading to temporal misalignment across views and reduced reconstruction quality. To address this challenge, a novel temporal alignment strategy is proposed for high-quality 4DGS reconstruction from unsynchronized multi-view videos. Our method features a coarse-to-fine alignment module that estimates and compensates for each camera's time shift. The method first determines a coarse, frame-level offset and then refines it to achieve sub-frame accuracy. This strategy can be integrated as a plug-and-play module into existing 4DGS frameworks, enhancing their robustness when handling asynchronous data. Experiments show that this approach effectively processes temporally misaligned videos and significantly enhances baseline methods.

Generative artificial intelligence chatbots are powerful tools that support users with disabilities across diverse contexts, including work, communication, and everyday life.Specifically within the neurodivergent population, numerous studies have illustrated increased use of LLM chatbots in applications for mental health, well-being, education, and professional support, along with their further potential to serve this community.Our work builds on this scholarship by investigating how neurodivergent individuals use Large Language Model-based chatbots in real-life settings.We conducted interviews with 15 neurodivergent participants to understand their expansive use of these chatbots, which we present as use case categories.We then designed diary studies to capture prompting strategies participants employ within these use cases while conversing with the Large Language Model-based chatbot Gemini.Based on early analysis of our conversational dataset, we showcase specific prompting queries from participants, highlight participant strategies to refine Gemini's responses, and identify Gemini's strengths and weaknesses during interaction.

Marginalized and underrepresented groups, including the disability community, are at risk of being harmed by bias and underserved by emerging technologies, such as Large Language Models (LLMs) and their downstream applications. While previous work has identified types of harm LLM technologies imposed on the disability community, there is a gap in research understanding the use cases and needs of people with disabilities when interacting with such technologies. To fill this knowledge gap, our study seeks to investigate real-life interactions between people with disabilities and LLM-based chatbots. Grounded in the Disability Justice Principle of following the leadership of the “the most impacted,” we interviewed 30 people with disabilities to learn about their current uses of chatbots. We analyzed 18 interviews and characterized distinct use cases, such as navigating social interactions, revising written work, and playing games, into nine broader categories. Our discussion contemplates how the presented chatbot use cases can serve as a foundation for further disability representation work with LLMs.