Please use this identifier to cite or link to this item:
|Title:||Smart receipt system - image processing, data analytics and server development||Authors:||Tham, Guo Bin||Keywords:||DRNTU::Engineering||Issue Date:||2018||Abstract:||The Smart Receipt system(SRS) was developed to create an easy way for users to keep track of daily expenditure. The SRS comprises of multiple sub systems which includes development of a mobile application, back-end server, and a database which uses the Google’s Firebase. The idea was to use techniques like Optical Character recognition(OCR), and pre-image processing to process receipt images and convert it to digital information so users can keep track of their expenditure with ease. Pre-image processing technique’s main aim was to improve image quality to improve OCR recognition rate. In addition, the use of Naïve Bayes classification to automatically classify receipt items into different types of categories for user to view how much they spend on each type of item. Lastly, the implementation of Rule association mining to identify user’s buying pattern and recommend different kind of items according to their buying habits. The SRS was deployed and tested within Nanyang Technological University by deploying it to Cyber Security Lab’s server. The mobile application was deployed to Google Play store for users to test and use the financial application. Lastly, further improvements can be made to improve the Naïve Bayes classifier by adding a self-learning module to add new categories to the classifier. Also, improvements can be made to the OCR engine to improve the recognition rate.||URI:||http://hdl.handle.net/10356/73899||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||SCSE Student Reports (FYP/IA/PA/PI)|
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.