Development of MiniXplorer: An image recognition mobile application using google machine learning kit and text-to-speech integration

Alamil, Brendiz Allain and Nazareno, Kristine and Santos, Zinnielle Ann and Chua, Joey and Ayo, Eliza (2025) Development of MiniXplorer: An image recognition mobile application using google machine learning kit and text-to-speech integration. World Journal of Advanced Research and Reviews, 25 (2). pp. 899-911. ISSN 2581-9615

This study presents the development and evaluation of MiniXplorer, a mobile learning application designed for children that integrates Google’s Machine Learning Kit (ML Kit) for real-time image recognition and Text-to-Speech (TTS) technology for auditory feedback. Employing a mixed-method descriptive developmental approach, the research combined interviews, surveys, observations, and data analysis to assess the application’s functionality, performance, and user experience holistically. Technical evaluations revealed MiniXplorer’s robust image processing capabilities across diverse parameters: it supported multiple file formats (notably .jpg), handled resolutions from low to high, and managed file sizes ranging from <1MB to >20MB. The application demonstrated high accuracy in natural lighting conditions, recognizing colorful objects, adapting to orientations (front/side views), and addressing edge cases such as partial obstructions or complex backgrounds. Performance testing confirmed consistent operation under varying noise levels and compatibility with modern Android OS versions. Security analysis identified minor vulnerabilities in the AndroidManifest.xml configuration, specifically the allowBackup and debuggable settings, which posed risks of sensitive data exposure. These were addressed to mitigate potential breaches. User evaluations aligned with ISO 25010 standards highlighted strong positive feedback, particularly praising the app’s object recognition accuracy, intuitive auditory feedback, smooth performance, and cross-device portability. Participants emphasized its usability and reliability as an educational tool for children. The study underscores the efficacy of integrating ML Kit and Flutter TTS in developing child-centric image recognition applications, successfully meeting functional, performance, security, usability, reliability, and portability criteria. MiniXplorer represents a scalable model for enhancing interactive learning through adaptive mobile technologies, demonstrating promise for broader educational applications.