Skip to content

gjentig4/sumthing_ML_model

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

Image Classifier for Sumthing.org

Introduction

Welcome to the Image Classifier project! This Jupyter Notebook details a machine learning solution developed for Sumthing.org. The NGO, dedicated to environmental conservation, needed a tool to automatically sort batches of images into two categories: 'trees' and 'plastic'. This classification aids in documenting and showcasing their efforts in tree planting and garbage collection to donors and the public.

Project Objective

  • Aim: To develop an automated image classifier for efficiently sorting images into 'trees' and 'plastic'.
  • Benefit: Streamlines image management and enhances the presentation of the NGO's environmental initiatives.

Approach

  • Method: Utilized TensorFlow and Keras to build a neural network, adopting the MobileNetV2 model with transfer learning.
  • Outcome: Achieved a remarkable 100% accuracy in image classification, showcasing the model's efficiency and reliability.

Application

  • The classifier processes images sourced from local storage or AWS S3 buckets.
  • Post-classification, the images are sorted into designated folders on AWS S3, simplifying access and documentation.

Impact

  • The model significantly aids in managing large image datasets.
  • Provides a visual record of the NGO's impactful work, boosting transparency and donor engagement.

Data Collection

  • Technique: In addition to the existing image dataset, I extended the data collection process through web scraping.
  • Tools Used: Utilized Python libraries like BeautifulSoup and Selenium to gather images from various online sources.
  • Purpose: This effort enriched the dataset, providing a broader range of images for training and enhancing the classifier's robustness. """

Conclusion

This project exemplifies the application of machine learning in environmental conservation, offering a practical solution to a real-world challenge faced by Sumthing.org.

Feel free to explore the code and reach out for any queries or collaborations on similar projects!

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published