Core Technologies

Vision-AI Model Design

oToBrite's vision-AI models are highly flexible and customizable for passenger cars, commercial vehicles, and two-wheelers. To maximize efficiency, oToBrite’s vision-AI models have been optimized for low-bit cross-platform integration, allowing them to run seamlessly on various hardware architectures, whether on high-performance computing platforms or resource-limited edge devices. Using a software IP licensing model, customers can easily apply the technology to meet their needs, enhancing business value. So far, over 400,000 vehicles have adopted this technology. Trained with deep learning algorithms and extensive labeled data, oToBrite's models accurately classify and detect objects, ensuring fast response times and low power consumption.

Learn more about Vision-AI models

AI Edge Computing System

Through precise design and optimization, our AI edge computing systems seamlessly integrate more than 10 heterogeneous cores and process data inputs from over 30 different sensors, including high-resolution camera modules, ultrasonic radars, millimeter-wave radars, GPS/IMU, and vehicle CAN signals. These sensor data are processed in real time, achieving high-precision environmental perception and decision-making, applied in autonomous driving and advanced driver assistance systems (ADAS). We have successfully mass-produced over 200,000 edge computing systems with advanced vision-AI models, which are widely used in commercial vehicles, passenger cars, and professional field autonomous robots, demonstrating their strong adaptability and stability across diverse markets.

Active Alignment Technology

The image quality of automotive cameras is vital for ADAS and autonomous driving, with clarity and consistency relying on precise autofocus control. oToBrite’s active alignment technology offers highly accurate, efficient solutions for high-end camera modules, enabling fast switching and flexible scalability to improve production efficiency. Developed over years of research, the five and three-axis alignment system surpasses traditional manual single-axis methods, significantly enhancing image clarity and consistency. oToBrite produces over 60 camera module specifications, each with 1 pixel-level offset precision, ensuring optimal performance across various applications. This independently developed technology meets the diverse customization needs of global markets.