Challenges

Limitations of Traditional AI CCTV

Traditional AI CCTV relies on pre-defined dataset-based models, which can only detect limited scenarios.
Applying new scenarios requires extensive data collection and labeling, resulting in significant time and cost.
Even when false alarms occur, users cannot correct them directly and must rely on vendor maintenance, which adds extra costs.
As a result, detection performance often falls short of expectations, and simple alarms without context are frequently ignored.

AI-powered Intelligent Safety Environment

EVA that Evolves through Conversation

EVA combines ML, LLM, and VLM technologies to instantly transform standard CCTV into AI-powered CCTV.
By leveraging various foundation models, users can easily define and apply scenarios through a conversational interface.
With natural language control and on-site customization features provided by the Agent, safety managers can set up and operate the system without AI expertise.

Applying Various Safety Scenarios

Detection objects and scenarios can be set for each camera via chat, and results are applied immediately.
By entering scenarios such as PPE detection, fall detection, or restricted area access, EVA efficiently performs object and event detection.
Detection results are delivered to Teams messenger along with images and evidence, enabling reliable response.

Improving Performance through Feedback

When false detections occur, on-site feedback can be incorporated to continuously improve detection performance.
False alarms may arise from misidentifying objects or confusing similar scenarios.
EVA refines detection logic by integrating user feedback with Multi-modal RAG, enabling more precise judgments.

EVA that Continuously Improves

EVA applies the most suitable ML, LLM, and VLM models depending on the environment and camera conditions to ensure optimal performance.
Its App and Agent architecture efficiently manages multiple foundation models to optimize costs.
All models and logic are continuously updated through the Mellerikat platform, ensuring EVA evolves consistently over time.

Cisco Meraki Integration

Scalable AI Safety Environment

Cisco Meraki is a cloud-based managed VMS service that makes camera expansion simple.
EVA easily integrates with Meraki cameras in both cloud and on-premise server environments to provide AI-powered safety solutions.
By combining the scalability of Cisco Meraki with EVA’s efficient architecture, organizations can accelerate cost-effective AI safety deployment.