Form Menu Intelligence

Overview

FORM, a leading provider of retail execution software since 2001, needed automated extraction of beverage menu data from restaurant and bar menu cards. Manual collection was slow, inconsistent, and could not scale. We built a machine learning pipeline that uses OCR, convolutional and recurrent neural networks, and LSTM segmentation to read menu images and structure wine, spirits, cocktail, and ingredient data for sales and market intelligence teams.

Project Highlights

• OCR with CNN/RNN deep learning models
• LSTM-based menu text segmentation
• Automated categorization of drinks, beer, wine, and meal sections

APPROACH & SOLUTION

Combined computer vision, natural language processing, and object detection models to replace manual menu surveys with structured data pipelines feeding business intelligence workflows