Desenvolvimento de aplicativo Android para treinamento de pronúncia integrada ao Anki

Language learning on mobile devices has been growing more and more, and tools to increase vocabulary and practice pronunciation help in the process. This document presents the general context on pronunciation assessment through software, on the use of artificial intelligence for the generation of flashcards and image recognition using OCR and computer vision. The objective of the work is the development of an Android application that integrates the functionalities of vocabulary import by recognition of highlighted words using OpenCV, creation of flashcards integrated with AnkiDroid using the Google Gemini API and pronunciation assessment with the available cards using Azure Speech Services. The result obtained was a functional open source application published on GitHub available for versions of Android 10 or higher. The work highlights the technologies used and the techniques for implementation and validation of the quality of the services used in the functionalities created.