10 datasets found

The dataset is used for lane detection in highway and city scenes.

The comma2k19 dataset is used to evaluate the robustness of lane detection models under physical-world adversarial attacks in autonomous driving.

The TuSimple Lane Detection Challenge dataset is used to evaluate the robustness of lane detection models under physical-world adversarial attacks in autonomous driving.

The CULane dataset is a widely used benchmark for lane detection. It contains 100,000 images, divided into three subsets: training set, validation set, and test set.

A hybrid spatial-temporal deep learning architecture for lane detection

A dataset of 2000 RGB images for lane detection using SegNet architecture.

The LLAMAS dataset contains more than 100,000 high-resolution images (1276 × 717), collected primarily from highway scenarios.

The TuSimple dataset contains 6408 annotated high-resolution (1280 × 720) images taken from video footage, each containing 2 to 4 lanes with clear markings.

The Curvelanes dataset contains more than 100,000 high-resolution images (1276 × 717), collected primarily from highway scenarios.

The CULane dataset contains clips from a video of more than 55 hours in length, which was collected by drivers in Beijing city streets.