Inspiration
Our project was inspired when we noticed during the first day of the hackathon how the garbage bin was overflowing with different kinds of waste after the snacks provided. This ignited within us a collective passion for environmental sustainability and a desire to make a meaningful impact on waste management and recycling. As individuals, we were concerned about the growing environmental issues caused by improper waste disposal and the lack of accessible tools to help people make informed choices about waste classification.
What it does
EcoSort is an image classification app that empowers you to make a positive impact on the environment. With just a snap, our app identifies and categorizes recyclable and organic waste items in seconds. Simply point your camera at an object, and EcoSort will instantly tell you whether it's recyclable, organic, or garbage waste.
How we built it
Dart , Flutter and Flask: We utilized the Dart programming language and Flutter for building the mobile app. We use flask to make an efficient server to support the app's functionality.
Azure Custom Vision: We integrated Azure Custom Vision, a powerful image classification AI service, for waste classification. Python was used to link our app's output data to the Azure Custom Vision cloud service.
Challenges we ran into
While developing EcoSort, we encountered several challenges:
Training the image classification model: Achieving high accuracy in waste classification required extensive training, time and data preprocessing to account for various lighting conditions and object orientations.
Integration with Azure Custom Vision: Linking our app's output data to the Azure Custom Vision cloud service required careful coding and configuration.
Despite these challenges, our commitment to creating a valuable and user-friendly tool for waste classification kept us motivated and focused on our mission.
Accomplishments that we're proud of
Successfully developing an app that can predict and classify waste using image recognition technology within the limited timeframe of the hackathon.
Raising awareness about waste management and environmental sustainability through technology.
What we learned
Throughout the development of EcoSort, we gained valuable insights:
The advantage of using Azure Custom Vision to can train object detection and classification models in the Azure cloud.
The significance of data accuracy and continuous improvement in training machine learning models.
The potential of technology to address real-world environmental challenges and promote positive behavioral change.
What's next for Waste Management
Enhance the accuracy of waste classification by continually refining and training our image recognition model.
Include educational resources within the app to further educate users about sustainable living and waste reduction.
Log in or sign up for Devpost to join the conversation.