Breadtec: An assistant detection app for common filipino bread varieties using yolo11/
Malcolm Xavier L. Dizon, Benj Anne T. Mameles, Rhea Miguela M. Manansala, Hana Joy Q. Marinduque, and Juan Gabriel Q. Segovia.--
- Manila: Technological University of the Philippines, 2025.
- 119pages: 29cm.
Bachelor's thesis
College of Science.--
Includes bibliographic references and index.
Filipino bread varieties represent an important aspect of the country's culinary heritage, yet many people struggle to identify specific types. This study developed BreadTec, a mobile application that uses machine learning to detect and classify common Filipino bread varieties from images. The research followed the Cross-Industry Standard Process for Data Mining (CRISP-DM) methodology for model development and Agile methodology for mobile application development. A YOLO11s object detection model was trained on a dataset of Filipino bread images collected from local bakeries and web sources, then deployed to Google Cloud Run for real-time inference. The Android application was developed using Android Studio and integrated with Firebase for data storage. The trained model achieved a mean Average Precision (mAP@0.5) of 0.909, with detection accuracy exceeding 95% under optimal conditions and above 85% under challenging conditions. User evaluation using ISO 25010 quality characteristics with 40 respondents showed an overall mean score of 3.37 out of 4.0, surpassing target scores and indicating strong user satisfaction across all measured quality attributes. The study demonstrates the technical feasibility of using deep learning for cultural food recognition applications. BreadTec effectively combines technological functionality with cultural education, providing a digital tool for preserving Filipino culinary heritage while potentially supporting local bakeries through enhanced consumer knowledge of traditional bread varieties.