Development of image classification apps using Android Studio

Ong, Chee Teng (2025) Development of image classification apps using Android Studio. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Development of image classification apps using Android Studio.pdf - Submitted Version Download (3MB)

Abstract

Image classification has emerged as a crucial task in the field of computer vision, facilitating numerous applications across various domains. It known in this area of study includes the training of sound and reliable machine learning models to classify the images accordingly into the categories defined. The problem is image classification application require many resources to be used and some challenge about reliability of currently available image classification apps. Those applications cannot classify the object very detailed. For the objectives of the project, the application needs to provide easy to use UI, detect more specific items and improve the accuracy of result for the image classification. To complete the project, the android studio was used to create an android application and TensorFlow Lite were trained using machine learning that import into the android app. After that, the application can recognize the image and give the result. The apps should be a simple UI and easy to use. Since this project is related to image classification, the analysis part will use accuracy testing to evaluate this app. The results show that developed apps have successfully detected the correct image up to 90% accuracy. The higher the accuracy, the greater the performance of model implement in the image classification app. This project presents how mobile apps can use machine learning to do something that can bring advanced features like real time classification. The image classification app can be used in many domains like education to help enthusiasts recognise species or serve as an educational tool for student and environmental conservation. In business, it could be adapted into specialized software for industries requiring automated identification.

Actions (login required)

View Item