Skip to content

Akashravic/Smart-Waste-Management-ML

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

4 Commits
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 

Repository files navigation

โ™ป๏ธ AI-Powered Urban Waste Classifier

An Automated Segregation System for Self-Sustainable Smart Cities


๐Ÿ“Œ Project Overview

Proper waste segregation is a cornerstone of sustainable urban development. This project utilizes Deep Learning and Computer Vision to automate the identification of household waste, categorizing materials to reduce recycling contamination.

The system was developed using Transfer Learning on the InceptionV3 architecture, achieving significant improvements in generalization compared to baseline lightweight models.


๐Ÿš€ Key Features

  • Multi-Scale Detection: Uses Inception modules to identify waste items of various shapes and sizes.
  • Fine-Tuned Accuracy: Optimized through partial unfreezing of deep layers to specialize in waste textures.
  • Interactive UI: A real-time web dashboard for image uploading and classification.
  • Smart City Logic: Integrated decision-making to label items as "Recyclable" or "General Trash."

๐Ÿ“Š Technical Benchmarks

I performed architecture benchmarking to determine the best model for this use case.

Metric MobileNetV2 (Baseline) InceptionV3 (Final)
Training Accuracy 80.2% 75.9%
Validation Accuracy 47.9% 70.0%
Key Advantage Resource Efficient Superior Generalization

๐Ÿ› ๏ธ Tech Stack

  • Language: Python 3.10
  • Deep Learning: TensorFlow 2.15+, Keras (Legacy Support via tf-keras)
  • Architecture: InceptionV3
  • Frontend: Streamlit
  • Deployment: Virtual Environments (venv)

About

An AI-powered Waste Classification system for Smart Cities using InceptionV3 Transfer Learning and Streamlit.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages