Applying Video Understanding and RAG in Surveillance

Video surveillance cameras are leveraged in the public sector for a variety of use cases like public safety and law enforcement, urban planning, public health, and more. Generative AI, using retrieval augmented generation (RAG) and pre-trained large language models (LLMs), can be used to analyze video data from these public sources. These technologies can enable the summarization and querying of this data to be much faster, more accurate, and more cost efficient than manual analysis.  

In this TechTalks session, we’ll explore RAG, a method of improving the accuracy and relevance of inference capability of LLMs, the evolution of multimodal LLMs, and how they can summarize and extract insights close to real time.

Join us as our experts discuss: 

• How multimodal LLMs makes extraction of insights from video much simpler and faster

• Overall trends shifting from text to video for better precision of answers

• Capacity and performance requirements for video data sets 

• Why camera companies stand to benefit from the advent of RAG