Antares Vision Group, an Italian-based company that provides product and supply chain digitalization, traceability, and inspection, recently announced the launch of AI-GO, an artificial intelligence-driven visual inspection platform.
The system was created in collaboration with Oròbix, an AI service company within Antares Vision Group. It’s designed for faster, more accurate, and flexible inspection processes, according to the company, which says it builds on its existing experience in quality control technologies.
Utilizing AI for vision inspection
One of AI-GO’s core features is improved accuracy. When compared to traditional systems, Antares says AI-GO can detect chromatic impurities, glass fragments, cosmetic defects, or foreign bodies.
As a result, the system can reduce false rejects while increasing the overall quality of production.
“Artificial intelligence has become a fundamental strategic lever for industrial competitiveness: integrated in a secure, transparent and verifiable way, it enables us to offer our customers better performing inspection machines, able to improve product quality, reduce false rejects, and increase production efficiency,” said Gianluca Mazzantini, CEO of Antares Vision Group, in a statement.
Built-in learning technologies
The system is integrated with learning technologies to adapt to various formats and products. This helps reduce set-up time and improves operational efficiency of production lines.
Antares said AI-GO can help a system understand, adapt, and improve to create a foundation for more intelligent and sustainable manufacturing.
“With solutions like AI-GO, we are bringing an intelligence to production processes able to understand real data and dynamically adapt itself to the variability frequently seen in production,” said Pietro Rota, CEO of Oròbix.
Antares says its goal with AI-GO is to guide companies toward advanced production models, including those where quality is obtained through objective data, predictive analyses, and digital process management.
Inside the AI-Go System
AI-GO consists of two components, according to Antares. The first is AI-GO Studio, a cloud-based software that gathers images and trains models.
The second is AI-GO Runtime, an edge module that enables autonomous execution of models directly into production lines. The system also works in offline settings.
With these capabilities, Antares says AI-GO can support operations and be implemented in complex industrial environments.
Upcoming features and packaging uses
In the future, Antares has plans to integrate Vision Language Models into the AI-GO platform. With this, the system can become a more seamless AI-enabled solution that builds on visual inspection and quality control, the company said.
“Vision Language Models are innovative models that combine the understanding of images and language, enables complex tasks to be performed through simple instructions in natural language,” said Luca Antiga, CTO of Oròbix. “We are working to render AI-GO increasingly intuitive and powerful: the native integration of VLMs will enable quality controls to be set up faster and in a more accessible way, even without technical skills.”
In packaging, AI-GO can be deployed for various needs such as seal inspection, label and print verification, foreign object detection, and cosmetic defect detection.
To learn more about AI-GO, click here.
