Asilla Launches Full-Scale Rollout of VLM-Integrated "AI Security asilla" for Real-Time Video Analytics

#

News

Achieve "total video analysis" by integrating Physical AI on the same hardware through large-scale parallel streaming processing with NVIDIA DeepStream and inference efficiency improvements with NVIDIA Triton Inference Server and NVIDIA TensorRT.

Asilla Inc. (Head Office: Machida City, Tokyo; CEO: Tsuyoshi Onoue, hereafter "Asilla"), launches a full-scale rollout of its v"Toward a Safe and Comfortable World Through the Power of Technology," vision with “AI Security asilla” platform.

AI Security asilla” is a platform that integrates vision language model (VLM) situational awareness into a large-scale real-time streaming analysis platform that processes numerous camera images in parallel in real time on the same hardware. Built with NVIDIA Metropolis libraries for video intelligence, this platform enables comprehensive video analysis that handles everything from detection to semantic reasoning on a single platform. The platform is used in large-scale facilities such as commercial facilities, office buildings, airports, and railways.

Background of Development

In video analytics, increasing the number of connected cameras brings processing load and infrastructure costs. Plus, there is a growing need for complex situational understanding that traditional detection models alone cannot fully answer.

VLMs are advanced AI models that excel at situational understanding and provide rich contextual information but require more computational resources due to its integrated multimodal processing of video and language. The challenge is to balance the efficiency of large-scale parallel processing with the inference efficiency of running VLMs without increasing resources.

Asilla solved this challenge by leveraging theNVIDIA Metropolis libraries and accelerated compute.

Technical Configuration

This platform is built on NVIDIA Metropolis and efficiently executes inference for each AI model (behavioral recognition model and VLM) on a massive parallel streaming pipeline. This enables a configuration that runs detection systems and VLMs in parallel within the same hardware resource range.

Features of this platform

Massive Parallel Streaming Processing: Utilizes NVIDIA DeepStream to process multiple real-time camera streams simultaneously at scale.

Streamlining AI Inference: NVIDIA Triton Inference Server and NVIDIA TensorRT optimize inference across multiple AI models, maximizing throughput and hardware utilization on NVIDIA RTX Pro Blackwell GPU.

VLM Integration with the Same Hardware: By utilizing the resources available from improved inference efficiency, VLMs are installed on the same hardware as detection system models. In addition to conventional detection, it provides integrated understanding of the meaning of the video.

Total Video Analysis: Provides large-scale parallel streaming processing, including VLM, as a single product to reduce implementation and operational costs.

Future Developments

Going forward,, Asilla will further expand its platform with new use cases in marketing and caregiving. Asilla is also exploring the use of NVIDIA Cosmos and NVIDIA Metropolis skills to accelerate the integration of more agentic capabilities such as fusion search. At the same time, we will continuously improve the accuracy of anomaly detection and understanding, as well as optimize models tailored to each site.

Comments from Stakeholders

[Masahiro Wakasa, Director and CTO of Ajira Inc.]

We have been able to deploy the insights we have gained from behavioral recognition AI at scale and efficiently on the NVIDIA Metropolis foundation. By integrating VLM into the same hardware, we can provide customers with 'situation-intelligent' video analysis that goes beyond mere detection, at a realistic cost.

Running VLMs in real time against numerous camera footage is not easy, and this configuration is a major step beyond that barrier. Using this foundation as a foothold, we will further expand the possibilities of video analysis.

Related Press Releases

Ajira, a video analytics AI company, to fully roll out its proprietary VLM "AsillaVision" at major domestic facilities

https://jp.asilla.com/post/news-asillavision-20260624

Asilla selected for the national project "GENIAC" implemented by the Ministry of Economy, Trade and Industry and NEDO

https://jp.asilla.com/post/news-geniac-20260604

About the Provision

Full-scale deployment of "AI Security asilla," has started across multiple locations. . For more details about implementation, please contact us at the inquiry address below.

*NVIDIA, NVIDIA Metropolis, DeepStream, TensorRT, and Triton are trademarks or registered trademarks of NVIDIA Corporation in the United States and other countries.

What is "AI Security asilla"?

"AI Security asilla" is a system in which AI analyzes footage from existing security cameras 24 hours a day, 365 days a year, instantly detecting abnormal behaviors such as violence, falls, and intrusion, as well as behaviors requiring attention such as wandering, crowding, and signs of poor physical condition. Amid an increasingly serious shortage of security personnel, it captures anomalies that are easily overlooked by human monitoring and immediately notifies security staff and administrators. Because existing cameras can be used as-is, no capital investment is required, making it a next-generation security solution that maintains a high level of safety even with limited staff.