Litterati envisions a world free from litter by harnessing the power of data and community action.
Industry
Startups & Software
Teams & Services
Back-End Engineering, Machine Learning, Generative AI, Large Language Models (LLMs), Cloud Architecture
Tech & Tools
AWS (S3, Lambda, DynamoDB, Bedrock, Rekognition), Python, SageMaker, Image Annotation
Key Data Points
The Vision
Litterati envisions a world free from litter by harnessing the power of data and community action. By making litter visible and measurable, they empower individuals, organizations, and governments to drive impactful change.
The Goal
Analyze helps cities, businesses, and activists quantify and understand litter trends, pinpoint sources of pollution, and develop data-driven strategies to prevent waste at the source.
The Challenge
Litter has been an unstructured and difficult-to-quantify problem, making it hard for governments, NGOs, and businesses to take strategic action. Pre-existing manual data collection methods were inefficient, lacking the precision and scalability needed for real impact.
The Solution
Protagona’s AI-powered solution refines Litterati’s approach by integrating an automated annotation system that enhances ML-generated bounding boxes for litter detection, ensuring high-confidence classifications. This system:
- Uses AWS Rekognition and an LLM jury system to generate and refine litter identification.
- Processes and stores structured litter data in DynamoDB, enabling real-time analysis.
- Routes low-confidence cases for human review, ensuring accuracy while minimizing manual effort.