Lane Segmentation

🚀 This project is built based on InSPyReNet (ACCV 2022). Please refer to the original repository for training and other details.

1. Create environment

• Create conda environment with following command conda create -y -n lane python
• Activate environment with following command conda activate lane
• Install requirements with following command pip install -r requirements.txt

2. Preparation

Pre-trained Checkpoint

• Download ImageNet pre-trained checkpoint for backbone network from Link
• Download checkpoint from Link
• Move file as follows ./snapshots/HighwayLane/latest.pth. Create folder if needed.

Dataset

• Train
• Test

3. Inference

• Prepare your image folder
• python run/Inference.py --source [IMAGE_FOLDER_DIR]

Performance - KAIST Highway Dataset

• Maximum F1 Score: 94.8
• Maximum IoU: 88.5
• Throughput: 43 fps
• GPU Mem Usage: 1.5 GB