UTAR Institutional Repository

Skin Lesion Detection Using Deep Neural Network By Smart Handheld Devices

Tan, Hou Ren (2020) Skin Lesion Detection Using Deep Neural Network By Smart Handheld Devices. Final Year Project, UTAR.

[img]
Preview
PDF
Download (2308Kb) | Preview

    Abstract

    Early detection of malignant skin lesions improves patient survival rates. Conventional self-detection method for public possess subjectivity, inaccuracy, and require experience. The goal of this project is to develop an Android based mobile application with object detection deep learning integration that allows global users to perform malignant skin lesions self-detection easily using a smartphone, for overcoming the limitations of the conventional method. Transfer Learning has been performed on various object detection models using ISIC skin lesions dataset with TensorFlow Object Detection API. The selected object detection model is SSD MobileNet V2 with 93.9% of evaluation accuracy after training due to its lightweight architecture therefore suitable for smartphone integration. The selected model has surpassed existing classification model in terms of accuracy after validation with a new dataset. A mobile application has been developed successfully with Android Studio. The trained object detection model successfully integrated into the mobile application using Firebase ML Kit and has achieved low detection time on smartphones. The mobile application has been proven to be compatible with various Android versions and screen sizes after tested with 7 different smartphones using Firebase Test Lab.

    Item Type: Final Year Project / Dissertation / Thesis (Final Year Project)
    Subjects: R Medicine > R Medicine (General)
    Divisions: Lee Kong Chian Faculty of Engineering and Science > Bachelor of Engineering (Hons) Biomedical Engineering
    Depositing User: Sg Long Library
    Date Deposited: 18 Aug 2021 20:15
    Last Modified: 18 Aug 2021 20:15
    URI: http://eprints.utar.edu.my/id/eprint/4225

    Actions (login required)

    View Item