Aplicativo móvel de gestão de gasto auxiliado por visão computacional

Abstract

There are several applications to control your spending, both for mobile devices and for desktop applications, but most of them require that data of income and expenditure are entered by the user, which can be uncomfortable and time-consuming, especially when the activity is performed through a mobile device such as a mobile phone. The need to keep track of spending and the need to spend less time with activities that can be carried out or facilitated by computational resources, led to the objective of this work, being this the development of an application that allows automatic registration of expenses by analysing the image of the tax coupon using an OCR system for the recognition of characters and machine learning techniques and neural networks to help reduce the time to register and facilitate the use of the application. In the application, the concepts of image processing and machine learning were developed and implemented, thereby improving the user experience with the application and creating a machine learning capable of recognizing the category of a tax coupon through the words contained in it.