Polylanenet++: enhancing the polynomial regression lane detection based on spatio-temporal fusion

Deep learning has made significant progress in lane detection across various public datasets, with models, such as PolyLaneNet, being computationally efficient. However, these models have limited spatial generalization capabilities, which ultimately lead to decreased accuracy. To address this issue, we propose a polynomial regression-based deep learning model that enhances spatial generalization and incorporates temporal information to improve the accuracy. Our model has been tested on public datasets, such as TuSimple and VIL100, and the results show that it outperforms PolyLaneNet and achieves state-of-the-art results. Incorporation of temporal information is also advantageous. Overall, our proposed framework offers improved accuracy and practicality in real-time applications.