Image Segmentation Model | People Segmentation with DeepLabV3-ResNet50
🧍 People Segmentation using DeepLabV3-ResNet50
📌 Overview
This repository provides a complete pipeline for semantic segmentation of people in images using DeepLabV3 with a ResNet50 backbone, leveraging transfer learning with PyTorch.
The model is trained on a custom dataset with binary class labels: Background and Person.
📋 What's Included?
✅ Data Preprocessing & Augmentation
✅ DeepLabV3 Model Customization for Binary Classes
✅ Custom DataLoader for People Segmentation
✅ Training Loop with Validation & Loss Logging
✅ Cosine Annealing Learning Rate Scheduler
✅ Inference Pipeline with Mask Overlay Visualizations
🛠 Tech Stack
Python 🐍
PyTorch 🔥
Torchvision
Albumentations 🌈
Matplotlib 📊
OpenCV 📷
Google Colab ☁️
🔗 GitHub Link: https://github.com/NitinCVOrbit/Peopl...