Philippine Paper Bill Authenticator for the Blind Using Image Processing
Software as a Service
NoCategory
Thesis DesignClients
Techstack
Purpose
Blind users have no reliable way to verify cash by touch alone. This Mapua University device authenticates Philippine paper bills and announces their denomination, protecting visually impaired users from counterfeit money and short-changing.
Description
A Raspberry Pi device that photographs bills under ultraviolet light and runs them through a layered analysis pipeline: a TensorFlow convolutional neural network classifies authenticity, Canny edge detection and a Gaussian mixture model examine security features, and EasyOCR reads the serial number and denomination text. Results are spoken aloud so a blind user immediately knows the bill's value and validity, with a PyQt interface managing capture and testing. Model evaluation includes accuracy and confusion-matrix reporting.
