SeeFix – AI Civic Issue Detection 🚧
SeeFix is an AI-powered system that helps detect civic infrastructure issues like potholes and garbage using images.
Users can upload an image, and the AI model predicts whether the issue is relevant or not.
This project combines:
- Computer Vision
- Backend API
- Real-world civic problem solving
Project Workflow
User uploads image
→ Backend API processes request
→ AI model analyzes image
→ Prediction returned (pothole / garbage / normal)
System Architecture
Frontend (Upload Image)
↓
Flask Backend API
↓
Image Preprocessing
↓
AI Model (MobileNetV2 Transfer Learning)
↓
Prediction (pothole / garbage / normal)
Features
- Detect potholes and garbage using AI
- Image classification model
- Backend API for predictions
- Transfer learning using MobileNetV2
- Scalable smart city application idea
Tech Stack
Machine Learning
- TensorFlow
- Keras
- MobileNetV2
Backend
- Flask
- Python
Tools
- Virtual Environment
- GitHub
- VS Code
Dataset Structure
dataset/ garbage/ normal/ potholes/
Each folder contains images of the respective class.
Classes used:
- Garbage
- Normal
- Potholes
Model Details
Base Model: MobileNetV2
Training Method: Transfer Learning
Input Size: 224 x 224
Output Classes: 3
The base model is frozen and custom layers are added for classification.