Skripsi

PENGENALAN KATA DALAM BAHASA ISYARAT INDONESIA (BISINDO) SECARA REAL-TIME MENGGUNAKAN METODE YOU ONLY LOOK ONCE (YOLOv5)

---

Penilaian

---

Fire disasters that can occur suddenly can affect victims who do not have the ability to hear, so victims use sign language to communicate with firefighters. Limited communication between firefighters and victims will require a lot of time in evacuating victims. With these emergency conditions, researchers developed software to recognize Katas in Indonesian Sign Language (BISINDO) in real-time using the YOLOv5, which is a deep learning-based object recognition method to facilitate officers in carrying out the evacuation process. The software was built using 9 combinations of epoch and batch size, namely batch size 16 epoch 50, batch size 24 epoch 50, batch size 32 epoch 50, batch size 16 epoch 250, batch size 24 epoch 250, batch size 32 epoch 250, batch size 16 epoch 500, batch size 24 epoch 500, batch size 24 epoch 500. The testing process is carried out in nonrealtime with object input testing data and real-time using a webcam camera. The best accuracy result is 88.9% in nonrealtime testing with a combination of batch size 32 epoch 500. The real-time testing process obtained the best average accuracy of 86% and an average response time of 0.104s with a combination of batch size 32 epoch 500

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2307002820	T99827	T1201172023	Central Library (Referens)	Available

Detail Information

Series Title

-

Call Number

T1201172023

Publisher

Indralaya : Jurusan Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2023

Collation

xiv, 126 hlm.; Ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

005.107

Content Type

-

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Pemrograman
Jurusan Teknik Informatika

Specific Detail Info

-

Statement of Responsibility

FIRA

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment