ATOM™ is Powering the Future of AI in Health and Traffic Monitoring Systems
As nations race to build robust infrastructures to meet the demands of advanced AI applications, Rebellions is leading South Korea’s K-Cloud Project, a flagship national initiative to build cutting-edge AI infrastructure. ATOM™ delivered outstanding results in the MLPerf benchmark and has been deployed in data centers to power a new generation of AI-driven services and applications. ATOM™ surpasses traditional GPUs in both performance and operational efficiency.
AI Food Recognition and Detection
Food technology is undergoing a revolution, driven by the integration of AI to deliver smarter and more precise solutions. ATOM™ powers an advanced AI-based food recognition and detection service, capable of identifying foods and ingredients, estimating calorie content, and offering nutritional guidance tailored to dietary needs. This technology could unlock significant potential across industries by enabling automation in restaurants, improved nutrition tracking in consumer applications, and personalized diet management in healthcare.
ATOM™ achieved nearly 2x the frame rate and over 2x the energy efficiency compared to NVIDIA’s GPU while running Superb AI’s fine-tuned detection model, which was built on YOLO-v6L and trained on over 110,000 food images.
Rebellions’ SDK facilitated seamless integration resulting in a highly efficient food analysis system with real-world applications in health monitoring, food waste reduction, and customized diet planning. By bringing AI closer to everyday life, this system demonstrates the tangible benefits of Rebellions’ advanced technology.
Smarter Roads with Laon Road – AI-Powered Traffic Management
As smart cities evolve, AI is transforming how traffic is monitored and managed, paving the way for safer and more efficient urban environments. Partnering with Laon Road, Rebellions’ ATOM™ supports AI-driven traffic anomaly detection and intervention.
The system leverages ATOM™’s capabilities to analyze traffic in real-time, combining AI-powered video detection with advanced traffic management. ATOM™ outperforms traditional GPUs with over 2x FPS/W efficiency (1.52 vs. 0.70) and more than 2x faster latency (11 ms vs. 24.11 ms) compared to NVIDIA’s GPU. This performance gap is critical for enabling traffic management at scale with large and continuous data streams.
By identifying and tracking vehicles and pedestrians in real-time, the system delivers actionable insights into traffic patterns and proactively addresses irregularities such as violations or accidents. Integrated with a robust urban management platform, it enables municipalities to optimize traffic flow, monitor intersection performance, and improve road safety. ATOM™ is well-suited for latency-critical applications, offering fast response times and high throughput. Its energy-efficient design makes it ideal for large-scale smart city deployments, enabling real-time adaptation to changing traffic conditions.
Why ATOM™ Leads the Way
Both projects highlight why Rebellions’ AI accelerators stand out:
- Developer-Centric Software Stack: ATOM™’s software ecosystem, including its SDK and model zoo, simplifies model optimization and accelerates deployment.
- Performance: ATOM™ consistently delivers industry-leading inference speeds and energy efficiency, critical for real-time and large-scale applications.
- Scalability: ATOM™ is built to support projects like K-Cloud, offering seamless scalability and adaptability for diverse applications.
From public health to urban traffic systems, ATOM™ not only drives innovation but also addresses real-world challenges with precision and reliability.
Looking Ahead
As AI reshapes industries, Rebellions chips are at the core of public AI infrastructure driving innovation. These projects mark just the beginning of our journey. From building national AI frameworks to enabling smart city innovations, we remain dedicated to delivering high-impact solutions.
Stay tuned as we continue to collaborate with partners to enable a smarter, more efficient future with AI.