Article Per Year (5 Year)
p-Index From 2019 - 2024
0.23
P-Index
This Author published in this journals
All Journal BULLET : Jurnal Multidisiplin Ilmu
JUANDRI
Unknown Affiliation
Humanities Languange, Linguistic, Communication & Media Law, Crime, Criminology & Criminal Justice Medicine & Pharmacology Nursing

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Pengenalan Warna Terhadap Objek Dengan Model Analisis Elemen Data Warna Gambar Berbasis Deep Neural Network JUANDRI; Nizirwan Anwar
BULLET : Jurnal Multidisiplin Ilmu Vol. 2 No. 1 (2023): BULLET : Jurnal Multidisiplin Ilmu
Publisher : CV. Multi Kreasi Media

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Object detection, which simulates human visual perception in finding the most significant objects in a scene, has been widely applied to a variety of computer vision tasks. Although various RGB-D-based object detection models with promising performance have been proposed over the past few years, a deep understanding of these models and challenges in this area is still lacking. Artificial intelligence provides new feasibility to control dexterous prostheses. The RGB-D image database is built based on four important grip patterns (cylinder, sphere, tripod and lateral). The sample images in the RGB-D data set were obtained on a wide variety of everyday objects with different sizes, shapes, postures, as well as different lighting conditions and camera positions. In order to use RGB images and depth images for feature fusion more effectively, this paper proposes three fusion models: RGB-D concat、RGB-D Ci-add and RGB-D Ci-concat. Our RGB-D fusion based approach significantly improves hand detection accuracy from 69.1 to 74.1 compared to one of the most advanced RGB based hand detectors. Existing RGB or D-based methods are unstable in unseen lighting conditions: in dark conditions, the accuracy of the RGB-based methods drops to 48.9, and in backlit conditions, the D-based methods accuracy drops to 28.3
Co-Authors Nizirwan Anwar