Core Technologies
Algorithm Framework
Based on deep learning technology, it covers the entire chain of intelligent driving, achieving full-stack self-research and development from perception, multi-modal fusion, high-precision positioning to planning and control, breaking through the technical bottlenecks of the industry. Support BEV (bird 's-eye view) pre-fusion perception and end-to-end algorithms, combined with a self-developed model training framework, to ensure the high reliability and real-time performance of the algorithm in complex scenarios, and adapt to the diverse needs of passenger cars, commercial vehicles, etc.
Schematic diagram of the deep neural network structure
NO.1，ICDAR 2015 Challenge 2 “Focused Scene Text”（2016）
NO.1，Cityscapes Semantic Segmentation（2017）
NO.1，Kitti Object Detection (Car)（2017）
NO.1，PASCAL VOC Semantic Pixel Labelling（2017）
NO.1，Large Scale 3D Human Activity Analysis Challenge in Depth Videos（2017）
感知網(wǎng)絡(luò)模型.mp4)
