MaybeShewill-CV/lanenet-lane-detection

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2562 stars · 896 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

An unofficial TensorFlow implementation of the LaneNet deep neural network for real-time lane detection, based on the IEEE IV paper 'Towards End-to-End Lane Detection: an Instance Segmentation Approach'. It combines binary semantic segmentation and instance segmentation with a discriminative loss to detect lane lines, with pretrained weights on the TuSimple dataset.

Use cases

• detect lane lines in driving images
• run real-time lane detection for autonomous driving research
• segment lanes using instance segmentation
• evaluate a lane detection model on the TuSimple dataset
• retrain or fine-tune LaneNet on custom lane datasets

When to choose

• you need a ready-to-use LaneNet implementation with pretrained TuSimple weights
• you are doing research or prototyping on lane detection with TensorFlow 1.x
• you want to reproduce results from the LaneNet instance segmentation paper

When to avoid

• you need PyTorch or TensorFlow 2.x support
• you need a production-grade, actively maintained lane detection pipeline
• you require real-time video input processing out of the box, since the input pipeline is not fully optimized

Facets

library · maturity maintenance

computer-vision image-processing deep-learning machine-learning computer-vision autonomous-vehicles deep-learning image-processing python lane-detection instance-segmentation semantic-segmentation tensorflow self-driving tusimple-dataset encoder-decoder linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="MaybeShewill-CV/lanenet-lane-detection")