Autonomous driving is becoming one of the leading industrial research areas.Therefore many automobile companies are coming up with semi to fully autonomousdriving solutions. Among these solutions, lane detection is one of the vitaldriver-assist features that play a crucial role in the decision-making processof the autonomous vehicle. A variety of solutions have been proposed to detectlanes on the road, which ranges from using hand-crafted features to thestate-of-the-art end-to-end trainable deep learning architectures. Most ofthese architectures are trained in a traffic constrained environment. In thispaper, we propose a novel solution to multi-lane detection, which outperformsstate of the art methods in terms of both accuracy and speed. To achieve this,we also offer a dataset with a more intuitive labeling scheme as compared toother benchmark datasets. Using our approach, we are able to obtain a lanesegmentation accuracy of 99.87% running at 54.53 fps (average).