KI blueprint for the video search and summary, which is now available for providing video analysis -ai agents in the entire industry

The Age of Video Analytics AI agent is here.

Video is one of the defining features of modern digital landscape that is over 50% of all global data traffic. Dominant in the media and increasingly important for companies in the entire industry. It is one of the largest and most ubiquitous data sources in the world. However, less than 1% is analyzed on knowledge.

Almost half of the global GDP comes from the physical industry, which includes energy for automobiles and electronics. In view of the concerns about the lack of work, the production of onshoring efforts and the increasing demand for automation, the AI ​​agents of video analyzes will play a decisive role than ever and help to bridge the physical and digital worlds.

In order to accelerate the development of these agents, Nvidia today makes the AI ​​draft for video search and summary (VSS), which is powered by the Nvidia Metropolis platform, generally available and provide developers to create and provide highly capable AI agents for analyzing large sums of real-time and archiving videos.

A wave of Vision AI agents and productivity assistants who are operated by Vision Language Models (VLMS) are online. These video analysis -ai agents combine powerful computer vision models with the skills of Super Intelligent Language models (LLMS) and enable companies to see, search and summarize enormous videos. By analyzing videos in real time or checking terabytes of recorded videos, the AI ​​agents of video analytics enable an unprecedented value and opportunities in a number of important industries.

Manufacturers and warehouses use AI agents to increase the safety and productivity of employees. For example, agents can help distribute forklifts and to position the employees for optimal efficiency. Intelligent cities use video analysis -ai agents to reduce traffic jams and increase security.

https://www.youtube.com/watch?v=pw8tl_bjnwa

A blueprint to create different fleets of video analyzes -ai agents

The VSS blueprint is based on the Nvidia Metropolis platform and is based on VLMS and LLMs such as Nvidia Vila and Nvidia Llama Nemotron, Nvidia Nemo Retriever Mikrodienst and Retrieval-Augmented generation (RAG)-a technology connects LLMS with LLMs.

The VSS Blueprint includes the NVIDIA AI Enterprise software platform, including NVIDIA NIM Microservices for VLMS, LLMS and Advanced Ai -Frameworks for RAG. With the VSS blueprint, users can summarize a video 100x faster than in real time. For example, a one -hour video can be summarized in text in less than a minute.

The VSS Blueprint offers a variety of powerful functions that offer a robust video understanding, performance and scalability.

This version introduces the extended hardware support, including the possibility of providing a single NVIDIA A100 or H100 GPU for smaller workloads, and offers greater flexibility in resource assignment. The blueprint can also be used on the edge of the Nvidia RTX 6000 Pro and Nvidia DGX Spark Computing platforms.

The VSS blueprint can process hundreds of live video streams or burst clips at the same time. In addition to the visual understanding, it offers audio transcription. Converting language into text adds the context -related depth in scenarios in which audio is critical. B. training videos, keynotes or team meetings.

Industry leaders use video analysis -ai agents to increase the business value

Everyone, from the world's leading manufacturers to intelligent cities and sports leagues, use the VSS blueprint to develop AI agents to optimize operations.

Pegatron, a leading company for electronics manufacturers, uses the VSS blueprint to study business procedures and train employees with best practice. The company also integrates the blueprint into its Pegaai platform so that companies can build AI agents in order to transform manufacturing processes.

These active ingredients can absorb and analyze massive videooluminas to enable extended functions such as automated monitoring, detection of anomaly, video search and incident reporting. Visual Analytics Agent from Pegatron can be used to understand operating procedures for the assembly of the printed circuit board and to identify when the actions are correct or incorrect. So far, the agents have reduced the work costs of Pegatron by 7% and the error rates by 67%.

Additional leading Taiwanese semiconductor and electronics manufacturers build AI agents and digital twins to optimize their planning and operating applications.

Kaohsiung City, Taiwan, uses a uniform AI application developed by his partner left vision to improve the reaction times of the incidents. Previously, the city departments such as waste management, transport and emergency reaction were isolated by a new infrastructure, which led to slow response times due to lack of access to critical information.

The AI-driven use of left-wing vision driven by the VSS blueprint offers agents that combine real-time video analyzes with generative AI in order not only to recognize visual elements, but also to understand and tell complex urban events such as floods or traffic accidents.

Left vision currently provides timely insights into 12 city departments and can be scaled from 30,000 city cameras to over 50,000 by 2026. These findings offer improved situation awareness and data -controlled decision -making in the city services and the reduction in response times in the incidents by up to 80%.

https://www.youtube.com/watch?v=1ocukybMPW4

The National Hockey League used the huge insights with the VSS blueprint to rationalize and accelerate vision ai workflows. It manages massive volumes of game material.

The NHL is positioned with the huge insightedgine to search for video by petabytes from video so that the calling up highlights and moments in the game is almost instant. AI-controlled agent workflows further improve the creation of content by automatically cutting, marking and putting together video content to facilitate access and use.

In the future, the league could possibly use real-time AI argumentation to enable tailor-made insights such as player statistics, strategy analyzes or fantasy recommendations that were generated dynamically during live games. This end-to-end automation could change the creation, curating and delivery of media and set a new standard for the production of AI-controlled sports content.

Siemens uses its industrial copilot for operations to support the employees of Fabrikfloors in the tasks of equipment, error treatment and performance optimization. This generative assistant with AI-operated assistant offers reponat answers to device errors with information on operating and document data.

The copilot was created with a merger of VSS components such as VLMS, LLMS and Nvidia Nemo Microservices. The industrial copilot has led to quick decision -making and reduced machine downtime. Siemens has reported an increase in productivity by 30%, with the potential of 50% achieving.

Supported by an expanding partner ecosystem that creates sophisticated AI agents

NVIDIA partners use the VSS blueprint to accelerate the creation of Agentic Ai Video -Analysis features for your work processes and to shorten the development time from months to weeks.

Superb AI, a leading provider of intelligent video analyzes, has set up a sophisticated airport operating project at Incheon Airport to shorten the waiting times of passengers within a few weeks. In Malaysia, the solution provider ITMAX builds advanced visual AI agents with the VSS blueprint for the city of Kuala Lumpur in order to improve the city's overall management and to shorten the reaction times of the incidents.

In the advertising sector, Pyler integrated the VSS blueprint into its brand security (AID) and AD -Targeting (AIM solutions) in just a few weeks. With the help and AIM, Samsung Electronics increased the advertising effectiveness with brand and product orientation, high-quality advertising placements. BYD recorded the 4-way increase in his display click by aiming contextual and positive content, while the Hana Financial Group exceeded several branded campaign goals.

Fingermark is the application provider of Eyecue, a real-time computer vision platform used by Quick Service restaurants. Fingermark adds the VSS blueprint in Eyecue to transform video material into clear, implementable knowledge into the waiting times of Drive-Thru, service bottlenecks and incidents in connection with employees on a scale.

Try the VSS Blueprint on Build.nvidia.com and read this technical blog for further details.