Hand target detection based on improved YOLOv5
by Zhu Xu; Jinbao Meng; Juanyan Fang
International Journal of Wireless and Mobile Computing (IJWMC), Vol. 25, No. 4, 2023

Abstract: With the growing maturity of deep learning-based target detection algorithms, their deployment in intelligent service robots for target detection has become popular nowadays, in order to improve the precision of real-time hand detection and recognition by intelligent service robots, enabling them to detect hands accurately in a variety of environments. This paper proposes a hand detection method based on improved YOLOv5 deep convolutional neural network. YOLOv5s is selected as the base target detection model, the SE attention module is added to the network neck detection layer to guide the model to pay more attention to the channel features of small target to improve the detection performance, and the detection layer is added to enhance the feature learning ability of the network for target regions. The loss function of the detection model is optimised according to the hand image features to improve the confidence of the prediction frame. The experimental results show that the proposed hand detection method based on the improved YOLOv5 deep convolutional neural network can achieve a precision of 99.02%, which is 6.54% better than the original YOLOv5.

Online publication date: Fri, 08-Dec-2023

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Wireless and Mobile Computing (IJWMC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com