Smart Waste Classifier

Welcome to the Waste Classification Assistant! Upload a photo of your waste item (JPG or PNG), and I'll help you identify its category and provide useful recycling tips.

Comment

Smart Waste Classifier ♻️🧠

With the growing global concern around sustainability and waste management, we were inspired to create a simple yet effective tool that promotes environmental awareness. We wanted to empower individuals—especially students and young learners—to understand what type of waste they’re Smart Waste Classifier is an AI-powered web app that allows users to upload an image of a waste item (in JPG or PNG format). The app then classifies the item into one of five categories: Plastic, Paper, Metal, Organic, or Glass. Along with the classification, it also provides:

  • A recycling tip relevant to that category
  • An optional fun fact or environmental impact related to the waste type

How we built it

We used AWS PartyRock, a no-code AI app builder, which made it easy to integrate generative AI capabilities without needing to write complex code or host a backend. The key components of the app include:

  • An image upload widget that supports JPG and PNG formats
  • A text generation AI model that classifies the image and identifies the waste type
  • A secondary AI widget that provides recycling tips and fun facts

  • A clean and accessible user interface for all ages

  • Image recognition without a trained dataset: Since PartyRock doesn’t support custom training, we had to rely on prompt engineering to guide the AI.

  • Handling image quality issues: The app struggles with blurry or ambiguous images.

  • Prompt tuning: It took multiple iterations to get the AI to consistently respond with accurate and helpful waste categories.

  • Built a complete, working AI tool with zero lines of code

  • Successfully demonstrated a socially impactful use of generative AI

  • Made the app accessible and easy to use for students and non-technical users

  • How to build and deploy AI tools using AWS PartyRock

  • How to design prompts to guide large language models and image understanding

  • The importance of user experience when building educational tools

  • Integrate a custom-trained image classification model (using Teachable Machine or TensorFlow Lite)

  • Add voice feedback or accessibility features

  • Expand to support more waste types and multi-item sorting

  • Host environmental leaderboards or challenges to gamify recycling habits

Built With

Share this project:

Updates