Currency Detection Application for Visually Impaired People using Tensorflow Lite
Ruchika Bhakhar, Riman Mandal, Archna Goyal, Ashok Kumar, Rahul Singh, Harsh Vardhan · 2024
The primary aim of the research endeavour is to delineate the process of recognizing the distinctive character- istics of Indian banknotes through the utilization of a mobile application. The team of individuals committed to this project has demonstrated exceptional attention to detail and precision in the development of an intricate system for the identification of Indian currency. This intricate undertaking encompassed an exhaustive and meticulous procedure of instructing the system through the utilization of sophisticated methodologies and algorithms, leveraging the cutting-edge technology facilitated by TensorFlow to its fullest extent. Following this intensive training phase, the system seamlessly transitioned into a more lightweight and efficient version known as TensorFlow Lite, without any disruptions or complications. By strategically employing Convolutional Neural Network (CNN) methodologies for image classification, a thorough and detailed examination of the image dataset was meticulously undertaken, leading to the proficient extraction of complex and nuanced characteristics that are indispensable for the precise recognition and classification of various denominations of Indian currency. The collective efforts and extensive research carried out have ultimately led to the culmination of a sophisticated currency detection model. This model has been enriched with remarkable proficiency in the ability to differentiate and categorize various denominations of Indian banknotes with an unparalleled level of accuracy and precision.