My 12 Month Long Project.

My 12 Month Long Project

TrashCan — An application to segregate waste using deep learning.

Well, technically, the software project was 9 months long, but as it was an academic year, let's go along with the title! I graduated with a Bachelor of Technology in Computer Science and Engineering this May(2023) and one of the only requirements was to work on a project for at least a semester and present the outcome for graduating.

I, along with 2 other teammates worked on TrashCan — Mobile Application to Segregate Garbage Automatically. In today’s day and age, using AI for social good for cleaning the environment (even if it's a very small change) seemed like a good idea for the 3 of us😄This gave us an opportunity to do something original instead of just plagiarizing some project from GitHub.

The Motivation🚀

So why would we require an App to help us segregate waste? Well in 2022, India ranks last among the 180 participating countries in the environmental performance index. This is due to poor handling of waste management and pollution

With the emergence of several government schemes such as “Swachh Bharat Abhiyan” and “Smart Cities Mission”, there has been a surge in both garbage production and its collection. But, there is zero effort to segregate the waste.

A major part of segregation is still dependent on manual labor which is slow and inefficient. Hence, we propose an automated way to segregate waste into recyclable and non-recyclable components using Deep Learning.

What’s New?🤔

• The majority of solutions that are currently available classify garbage into a small number of classes (2 to 6 classes at most), but we provide 12 different classes.
• This includes paper, cardboard, biological, metal, plastic, green, brown glass, white glass, trash, clothes, shoes, and batteries. It relies on live image capture for generating output: we’ve simultaneously introduced multi-class categorization
• Realtime cloud deployment of the ML model with a fully functional cross-platform mobile application

The Technical Aspect💻

All of the source code with in-depth information can be found on the GitHub Repository.

Naturally, this project was very tech-heavy but can be explained easily in brief :

• Deep Learning — XceptionNet Model, a CNN developed by Google Research modified using TensorFlow [used Kaggle GPU]
• FastAPI for developing the API connecting the Model and the Mobile Application. It was hosted on the Google Cloud Platform for real-time access along with the dataset of 15500 images
• Lastly, Flutter Framework for developing a seamless cross-platform mobile application encapsulating all the elements of the project

The Team🫂

Although I was friends with my teammates for the past 3 years, it was quite a unique and refreshing experience to work seriously on a project with them. There was a steep learning curve which was a bit frustrating at times, but the results made up for our shortcomings. I hope it was a great journey for my friends as well.

Meet the team:

• Abhigyan Singh — Twitter, GitHub, and LinkedIn
• Naman Jain — Twitter, GitHub, and LinkedIn
• Suvodeep Sinha (Me) — Twitter, GitHub, and LinkedIn

In Conclusion🖊️

The project, team, and my graduation were a success! I would recommend all people in university to build something with their group of friends! It is a once-in-a-lifetime opportunity to learn and grow.

If you have any queries about the project (GitHub), you can contact me or anyone of us. Do check out my profile and give other articles a read as well!

Cheers ✌

More content at PlainEnglish.io.

Sign up for our free weekly newsletter. Follow us on Twitter, LinkedIn, YouTube, and Discord.

---

My 12 Month Long Project. was originally published in Artificial Intelligence in Plain English on Medium, where people are continuing the conversation by highlighting and responding to this story.

https://ai.plainenglish.io/my-12-month-long-project-ab374c28bbf3?source=rss—-78d064101951—4
By: Suvodeep Sinha
Title: My 12 Month Long Project.
Sourced From: ai.plainenglish.io/my-12-month-long-project-ab374c28bbf3?source=rss—-78d064101951—4
Published Date: Thu, 29 Jun 2023 01:08:44 GMT