Syauqiy, Izzulhaq Fawwaz (2024) Perancangan Dan Pembuatan Aplikasi Android Klasifikais Jenis Luka Melalui Kamera Ponsel Menggunakan Metode Transfer Learning. Other thesis, Institut Teknologi Sepuluh Nopember.
Text
05111840000023-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (12MB) | Request a copy |
Abstract
Luka adalah keadaan rusaknya suatu jaringan tubuh yang menyebabkan hilangnya sebagian fungsi tubuh akibat trauma pada bagian tubuh tertentu. Penyebab trauma tersebut dapat berupa benda tumpul, benda tajam, atau permukaan yang kasar. Penyembuhan luka sendiri bergantung pada bagaimana luka tersebut dirawat, dan untuk merawat luka dengan tepat dibutuhkan pengetahuan tentang luka yang dialami. Aplikasi identifikasi jenis luka dibuat untuk mempermudah dan mempercepat pengenalan luka sehingga perawatan yang tepat dapat dilakukan. Memanfaatkan fitur kamera pada smartphone untuk mengambil gambar luka dan menggunakan artificial intelligence (AI) untuk menentukan jenis luka dari gambar yang diambil. Dataset yang digunakan mengalami pemrosesan guna mendapatkan region of interest (ROI) dari gambar untuk meningkatkan performa pelatihan model. Model machine learning dilatih menggunakan metode transfer learning untuk mengatasi masalah sedikitnya jumlah data yang tersedia. Model yang dilatih dengan dataset yang dilakukan ROI cropping berhasil mencapai akurasi sebesar 89%, sedangkan model yang dilatih dengan dataset tanpa ROI cropping mendapatkan akurasi sebesar 83%.
========================================================================================================================
A wound is a condition of damage to body tissue that causes the loss of some body functions due to trauma to certain parts of the body. The cause of the trauma can be a blunt object, sharp object, or rough surface. Wound healing itself depends on how the wound is treated, and to treat wounds properly requires knowledge about the wound experienced. The wound type identification application was created to make it easier and faster to recognize wounds so that appropriate treatment can be carried out. Utilize the camera feature on your smartphone to take pictures of wounds and use artificial intelligence (AI) to determine the type of wound from the images taken. The dataset undergoes processing to obtain regions of interest (ROI) from the images to improve the training performance of model. Machine learning models are trained using the transfer learning method to overcome the problem of the small amount of available data. The model trained on the dataset with ROI cropping achieved an accuracy of 89%, while the model trained on the dataset without ROI cropping achieved an accuracy of 83%.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | android, klasifikasi gambar, luka, machine learning, transfer learning, android, image classification, machine learning, transfer learning, wound. |
Subjects: | T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing |
Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
Depositing User: | Izzulhaq Fawwaz Syauqiy |
Date Deposited: | 06 Aug 2024 07:33 |
Last Modified: | 06 Aug 2024 07:33 |
URI: | http://repository.its.ac.id/id/eprint/111962 |
Actions (login required)
View Item |