Deep_Lane_Segmentation

KMU HCI Lab

Deep Lane Segmentation using SHG Module, Unet

---

Folders

Folder	Description
models	network architectures and parameters
datasets	define dataloader interfaces
criterions	define criterion interfaces
util	visulization utilities

---

Directory

D_Lane_Segmentation
├── src
│    ├── models
│    ├── input_video
│    ├── datasets   
│    ├── output_video
│    └── util
│
└── result
     ├── Unet_weight
     └── SHG_weight

Requirements

• python3(recommend) or python2
• pytorch==1.4.0
• torchvision
• opencv==3.3.1
• scipy, numpy, progress, protobuf
• joblib (for parallel processing data.)
• tqdm

Inference

Download Weight File

• Weight File Download SHG_weight, Unet_weight

mkdir result && cd result && tar -zxvf SHG_weight.tar.gz && tar -zxvf Unet_weight.tar.gz

Stacked Hourglass Network

python3 inference.py --netType stackedHGB --resume ../result/SHG_weight --nStack 7

Unet

python3 inference_unet.py --netType Unet --resume ../result/Unet_weight

Inference Video

KMU Self-Driving-Studio - SHG