Nota and Intel Launch 'Vision Agent x B70' Package, Processing 32 Video Channels Simultaneously
TECHWORLD ·
✦ AI Summary
Nota and Intel said on the 17th that they have launched the package product 'Nota Vision Agent x B70.'
The product consists of the video surveillance solution 'Nota Vision Agent (NVA)' and Intel GPU 'Arc Pro B70.'
In performance verification, one Intel B70 GPU processed 32 video channels simultaneously, and the average throughput per channel was 16.9 frames per second, exceeding Nota's actual ITS operating target of 15 frames per second.
Nota and Intel said on the 17th that they have launched the package product "Nota Vision Agent x B70" through a partnership. The product consists of the video surveillance solution "Nota Vision Agent (NVA)" and Intel GPU "Arc Pro B70."
The two companies focused on rising demand for applications in intelligent transportation systems (ITS), industrial safety, and selective surveillance at sites where video analytics needs are growing. In these fields, there has been a burden of separately selecting and optimizing hardware and software according to the number of CCTV channels, required performance, budget, and existing infrastructure.
Accordingly, Nota and Intel, as co-development partners, introduced a product that combines the Intel B70 GPU with Nota Vision Agent, which is optimized for computational architecture. The package aims to reduce that process, and its focus is on simplifying the introduction of video surveillance systems based on vision-language models (VLM).
Under the division of roles, Nota handled software optimization, while Intel provided technical support for the hardware and software environment. The product's processing performance is simultaneous handling of 32 video channels.
Performance verification was carried out based on the ITS operating environment that Nota established in Daejeon in 2025. The environment includes about 800 CCTV channels and 200 smart intersections. The verification reflected the entire surveillance workflow in actual operations, from video decompression to vehicle and person detection and tracking, and video conversion and transmission of analysis results.
Under these conditions, one Intel B70 GPU processed 32 video channels simultaneously. The total analysis throughput was 540 frames per second, and the average throughput per channel was 16.9 frames per second. This exceeded Nota's actual ITS operating target of 15 frames per second, which was used as the benchmark.
Nota Vision Agent is a video surveillance solution based on a combination of computer vision (CV) and VLM. The solution can analyze objects, behavior, and situational context in video and automatically generate reports on key content. It also applies model lightweighting and hardware optimization, and was designed to run in a variety of environments, including servers, cloud, mobile, and edge.
Using this product as a starting point, Nota and Intel will pursue joint customer discovery and also promote introduction consulting and on-site technical support together. Building on this, the two companies plan to work together from customer acquisition through adoption support and on-site support.
Based on this, the two companies plan to apply Nota's AI optimization technology to various Intel hardware products. They also plan to expand the scope of cooperation into industrial AI areas such as ITS, industrial safety, and selective surveillance.
Chae Myung-su, CEO of Nota, said the company aims to deliver the performance required by industrial sites through the combination of Nota's optimization technology and Intel hardware. He also said the company plans to expand options for building AI infrastructure suited to each customer's site. Michael Monson, product manager for Intel Arc Pro, emphasized the combination of Nota's optimization technology and the Arc Pro B70 GPU, and said Nota Vision Agent can be used in environments that offer strong performance and cost efficiency. He added that Intel will continue cooperation spanning hardware and software.
Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407100
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.