An exploratory computer vision project conducted for a global retail client during my time with Fellowship.AI. The company, which brings 2,000+ consumer packaged goods to market each year, wished to automate the QA process for their product packaging. The full project used computer vision tools and techniques to automate over 90% of the task, reducing processing time from hours or days to under ten minutes per product.
The files here consistute a demo of the exploratory work, conducting scene text detection and OCR using the Google Vision API, and quality control using functions based on the Levenshtein distance package.
Notebook demonstrating the text detection, OCR, and anlysis process. Uses functions from the following two scripts.
The core functions for text detection, interpretation, and comparison to groundtruth.
Calls core functions multiple times for improved detection accuracy.