Key Points Estimation and Point Instance Segmentation Approach for Lane Detection

Perception techniques for autonomous driving should he adaptive to various environments. In essential perception modules for traffic line detection, many conditions should be considered, such as a number of traffic lines and computing power of the target system. To address these problems, in this paper, we propose a traffic line detection method called Point Instance Network (PINet); the method is based on the key points estimation and instance segmentation approach. The PINet includes several hourglass models that are trained simultaneously with the same loss function. Therefore, the size of the trained models can be chosen according to the target environment's computing power. We cast a clustering problem of the predicted key points as an instance segmentation problem; the PINet can be trained regardless of the number of the traffic lines. The PINet achieves competitive accuracy and false positive on CULane and TuSimple datasets, popular public datasets for lane detection. Our code is available at https://github.com/koyeongmin/PINet_new

Automated summary

The proposed traffic line detection method, PINet, based on key points estimation and instance segmentation, is adaptive to various environments and computing power. PINet allows for choosing the size of trained models based on the target environment's computing power, achieving competitive accuracy and false positive rates on popular public datasets like CULane and TuSimple.