Trabalho em anais de evento

Comparative Evaluation of DNNs and Template Matching for GUI Validation on Android Devices

MYRRIA, ABDA; CASTRO, ALICE; MUNIZ, CAIQUE; BARBOSA, HERYCK; SILVA, ANDRÉ; CORDOVIL-JR, LUIZ; PORTY, ROGER; SOUTO, THIAGO

2026 • 2026 IEEE International Conference on Consumer Electronics (ICCE)

Resumo

The integration of Artificial Intelligence (AI) into software engineering has revolutionized the testing and validation processes for Graphical User Interfaces (GUIs), particularly in enhancing productivity, maintainability, scalability, security, and quality assurance (QA). This paper investigates the application of Deep Neural Networks (DNNs) to automate GUI testing within Android devices using computer vision algorithms. Specifically, we focus on automating the Home Host Test Case validation process from the Google Mobile Services Test Suite, a critical component in certifying Android devices for running proprietary Google applications. Our study evaluates the effectiveness of two DNN models — Faster R-CNN and SSD — in detecting objects within screenshots captured during testing, comparing their performance against the traditional Template Matching method. The comparative analysis highlights the strengths and limitations of each approach, considering the standardized nature of objects in these GUI tests. Through experimental evaluations, this research demonstrates how AI-driven automation can significantly enhance the efficiency and reliability of software QA processes for Android devices by reducing manual intervention and accelerating development cycles. Our findings underscore the transformative potential of DNNs in automating GUI validation tasks, contributing to a more agile and high-quality software development environment.

Autoria