Hand Gesture Identification System Using Convolutional Neural Networks

Recognition of hand movements is a key to conquering several difficulties and building warmth for human life. In an enormous number of applications, human actions and their significance are used in an array of applications to grasp the flexibility of machines. Sign language interpretation is one par...

Full description

Saved in:

Bibliographic Details
Published in:	2021 International Conference on Artificial Intelligence and Machine Vision (AIMV) pp. 1 - 7
Main Authors:	Arjaria, Siddhartha, Sahu, Riya, Agrawal, Sejal, Khare, Suyash, Agarwal, Yashi, Chaubey, Gyanendra
Format:	Conference Proceeding
Language:	English
Published:	IEEE 24-09-2021
Subjects:	and Hidden Markov Model (HMM) Assistive technologies Background Subtraction Convolutional Neural Network (CNN) Fault tolerance Fault tolerant systems Feature Extraction Gesture recognition Hidden Markov models Median and Gaussian Filter Neural networks Object segmentation Sign Language Speech impaired people
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Recognition of hand movements is a key to conquering several difficulties and building warmth for human life. In an enormous number of applications, human actions and their significance are used in an array of applications to grasp the flexibility of machines. Sign language interpretation is one particular area of interest. Following paper describes a practical and interactive procedure for hand gesture detection by making use of a Convolutional Neural Network. The techniques are suitably graded into various stages during the process, such as the data acquisition, pre-processing, segmentation, extraction of features, and classification. The different algorithms that have done their task at each location are elaborated, along with their merits. Challenges and limitations faced during the process are discussed. Overall, it is hoped that the analysis might provide a detailed introduction into the sector of machine-driven gesture and signing acknowledgment and further facilitation of future research efforts in this sector. The proposed methodology has been tested over the 8700 images, and it classifies the images with an approximate accuracy of above 95%.
DOI:	10.1109/AIMV53313.2021.9670906