.png){kind=icon}
.png){kind=icon}
Return
Object Detection Through Finger Count Using Deep Learning Techniques for HRI
.png){kind=icon}
PRE
AI
DOI：10.1007/978-3-031-93694-4_18.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The Human-Robot Interaction (HRI) system to help disabled and elderly community and improve their standard of living is still offers considerable potential for further exploration. This research work focuses on the automatic detection of objects through the automatic recognition of human finger count. The two Deep Learning (DL) techniques, MediaPipe and You Only Look Once (YOLO) v8 are employed for human finger count recognition and object detection, respectively. The study involves a mapping of finger count and corresponding object classes. A new dataset is created from scratch, encompassing four distinct object classes. Utilizing the Intel RealSense depth camera, the system provides position and depth information of the detected objects based on specific finger count. Simulation and experimental results indicate that the proposed approach outperforms recent related studies, demonstrating significant efficacy in developing an advanced HRI system.
Keywords:
Object detection
Finger counting
YOLO v8
MediaPipe
HRI system
Journal
C
IF:
0
Papers:
32
Citations:
0