In this talk, Dr. Jianquan Liu presents an industry perspective on the convergence of video analytics and generative AI. The talk begins with an overview of video analytics, covering advancements in action recognition, object tracking, human-object interactions, scene recognition, and behavioral pattern analysis. These technologies enable efficient extraction, retrieval, visualization, and summarization of video content. The presentation then explores the impact of generative AI, particularly large language models (LLMs), on video understanding. It discusses how LLMs enhance object recognition, semantic segmentation, action recognition, captioning, visual question answering, and storytelling. Dr. Liu provides industry case studies to illustrate these applications while also addressing limitations and challenges. The talk introduces NEC's narrative summarization framework, designed to tackle key challenges in video analytics. It concludes with a demonstration of "Video with LLM" technology, showcasing its practical application in automating traffic accident investigation reports. This presentation offers valuable insights into the current state and future potential of AI-driven video intelligence, bridging the gap between technical innovation and practical application for both industry professionals and general audiences.
Co-sponsored by: Oakland-East Bay Section Chapter
Speaker(s): Jianquan Liu ,
Agenda:
– 5:30 PM – 6:00 PM: Networking
– 6:00 PM – 7:30 PM: Presentation, Q&A
Sobrato Campus for Discovery and Innovation (SCDI) Room 1302, 500 El Camino Real, Santa Clara, California, United States, 95053
