Загрузка...

How to Build a Pothole Segmentation Model using U-Net and ResNet34 | Computer Vision

Pothole Image Segmentation using U-Net (ResNet34)
📌 Overview
This repository provides a complete pipeline for pothole segmentation from images using deep learning. It utilizes the U-Net architecture with a ResNet34 encoder backbone, implemented using PyTorch and segmentation-models-pytorch.

🚀 Features
Polygon annotation visualization ✅
Custom Dataset Loader for YOLO-format polygon annotations ✅
Pre-trained U-Net + ResNet34 encoder ✅
Cosine Annealing LR Scheduler ✅
Best model saving during training ✅
Segmentation predictions visualization ✅

🗂️ Dataset
📥 Download Dataset: https://drive.google.com/file/d/1d9qW_ctiCjKOiuA7-O6mX4cJgP3jcTDf/view?usp=sharing

#PotholeDetection #ImageSegmentation #ComputerVision #DeepLearning #UNet #ResNet34 #PyTorch #SemanticSegmentation #AI #MachineLearning #RoadSafety #AIForGood #OpenCV #VisionAI #SegmentationModels

Видео How to Build a Pothole Segmentation Model using U-Net and ResNet34 | Computer Vision канала CV orbit
Яндекс.Метрика

На информационно-развлекательном портале SALDA.WS применяются cookie-файлы. Нажимая кнопку Принять, вы подтверждаете свое согласие на их использование.

Об использовании CookiesПринять