Semantic road segmentation using U-Net with a ResNet34 encoder in PyTorch.
Deep learning framework for post-disaster road accessibility assessment using SegFormer-B2, graph analysis, and OpenStreetMap (OSM).
AI-powered road resilience system that detects and reconstructs occluded roads from satellite imagery using deep learning, enabling accurate mapping for disaster response, infrastructure planning, and smart mobility.
Real-time LIDAR-based Urban Road and Sidewalk detection for Autonomous Vehicles 🚗
Satellite road extraction pipeline: SegFormer-B2 semantic segmentation on DeepGlobe imagery (0.68 IoU) → morphological post-processing → A* pathfinding on predicted road networks. Includes a Gradio app and Docker image.
U-Net road segmentation pretrained on DeepGlobe and fine-tuned on UAVid (IoU 0.7168)
PyTorch implementation of SPIN Road Mapper: road segmentation from aerial images using spatial and interaction space graph reasoning on stacked hourglass networks.
Advanced autonomous driving perception suite comparing classical computer vision and deep learning based lane/road detection pipelines, including HybridNets ONNX multitask inference, U-Net road segmentation, curved lane tracking, and real-time GPU-accelerated autonomous road analysis.
Road damage segmentation using UNet++ & EfficientNet with 5-Fold Cross Validation, ensemble prediction, AMP training, and threshold tuning in PyTorch.
Semantic segmentation of road scenes using U-Net with ResNet50 on the CamVid dataset.
A modular, real-time ADAS stack that connects AI perception to physical robot control using ROS.
End-to-end real-time ADAS system combining perception, trajectory planning, and control. Runs efficiently on embedded edge devices using optimized ONNX inference pipelines.
Lane detection using U-Net (ResNet34) with Unsupervised Domain Adaptation (entropy minimization) on the CARLANE / MoLane dataset. Includes training pipeline, model exports, and evaluation dashboard.
基于YOLOv8的道路实例分割 | 像素级实线检测 | CUDA加速实时推理 | 软件杯竞赛作品
we introduce R2S100K---a large-scale dataset and benchmark for training and evaluation of road segmentation in challenging unstructured roadways.
An advanced, modular lane detection and road perception framework built with OpenCV, NumPy, and Gradio, engineered for real-time autonomous driving research. It features adaptive ROI mapping, probabilistic Hough transformation, EMA-smoothed lane tracking, and an interactive Gradio demo UI.
Road maps of the MIPT campus (Moscow Institute of Physics and Technology) and its surroundings are generated from satellite imagery using segmentation models (UNet and DeepLabV3) and road network graph reconstruction techniques.
U-Net-based semantic segmentation on images - pixel-level road and environment understanding.
ADAS system for Indian road scenarios using YOLOv8-Seg for road segmentation, object detection, and instance segmentation, trained on diverse real-world datasets with superpixel refinement and real-time inference capabilities.
A deep learning-based road segmentation model that leverages the U-Net architecture enhanced with attention mechanisms for better performance in detecting road features from satellite or street-view images.
Add a description, image, and links to the road-segmentation topic page so that developers can more easily learn about it.
To associate your repository with the road-segmentation topic, visit your repo's landing page and select "manage topics."