Model Development for Automatic Parking Sign Recognition – 90% Accuracy Including Recognizing Unstandardized Signs

Model-Development-for-Automatic-Parking-Sign-Recognition-–-90-Accuracy-Including-Recognizing-Unstandardized-Signs_preview-150
Client

The Client delivers innovations for the automotive & transportation industries, such as real-time parking and traffic information and solutions for AV’s safety testing and deployment.

Challenge
  • Building a web service with ML pipeline to automate data processing for parking sign recognition.

     

Benefits and Results
  • The algorithm identifies 85%+ of signs on the streets in the city, 90%+ identified correctly.

  • All the ML and data processing algorithms were implemented in a single web app with a detailed description of the model and project documentation.

  • The Client received a solution that automated parking sign detection and reduced the workload.