AI Dataset

Images paired with frame-corresponding point cloud data, can be utilized for sensor fusion in AI models, suitable for both late-fused and early-fused models.

Two traffic light pole cameras capture different perspectives of the same intersection, providing visual and spatial information essential for comprehensive scene analysis in AI applications.

Multi-camera views, including front, rear, side, and blind-spot cameras, are paired with LiDAR point cloud data with a coverage range of 100 meters.

This highly flexible dataset is suitable for training AI models in various ADAS applications, providing extensive environmental data that enhances the models' accuracy and reliability in real-world scenarios.

MVS stands for Motion Vector Sensor, which is a type of sensor that obtains motion information while capturing images.

Motion vectors can simplify the complexity of AI models, reducing power consumption and accelerating edge computing.