Machine Learning Approach to Classify Sparkling Wines

Strategic Insights -

Research conducted by the University of Wisconsin-Milwaukee has used a machine learning and artificial intelligence approach to classify types of sparkling wines and their containers using image data of bubble patterns. The study, supported by the Ministry of Science and Higher Education of the Russian Federation, employed a novel application of computer vision and machine learning techniques to analyze the characteristics of champagne. The results showed that the bubbles in different types of wine and glasses could be distinguished using machine learning techniques.

Key Takeaways:

  • The study used a machine learning approach to classify sparkling wines and their containers using image data of bubble patterns.
  • The research employed a computer vision analysis of video images and clustering using an artificial neural network approach.
  • The study integrated a segmentation neural network to filter out irrelevant frames and applied the Contrastive Language-Image Pre-Training (CLIP) neural network for feature embedding, followed by TabNet for classification.
  • The results showed that the bubbles in different types of wine and glasses could be distinguished using machine learning techniques.
  • The study demonstrated a novel application of machine learning and artificial intelligence for distinguishing champagne characteristics.
  • The research concluded that computer vision and machine learning analysis of ultrasound cavitation bubbles can be used to analyze carbonated liquids.
  • Dr. Michael Nosonovsky, a researcher from the University of Wisconsin-Milwaukee, was a co-author of the study.
  • Additional authors for the research include Timur Aliev, Ilya Korolev, Mikhail Yasnov, and Ekaterina V. Skorb.

Statistics:

  • The study analyzed two types of sparkling wines and two types of glasses.
  • The research used a computer vision analysis of video images to analyze the characteristics of champagne.
  • The study employed a machine learning approach to classify sparkling wines and their containers.
  • The results showed that the bubbles in different types of wine and glasses could be distinguished using machine learning techniques.
  • The study demonstrated a 95% accuracy rate in classifying sparkling wines and their containers using machine learning techniques.

Sources:

  • Rose or White, Glass or Plastic: Computer Vision and Machine Learning Study of Cavitation Bubbles In Sparkling Wines. RSC Advances, 2025;15(7):5151-5158.
  • NewsRx. Studies from University of Wisconsin - Milwaukee Describe New Findings in Machine Learning (Rose or White, Glass or Plastic: Computer Vision and Machine Learning Study of Cavitation Bubbles In Sparkling Wines). Computer Weekly News. May 14, 2025; p 744.