Title: The evolution of military operations: artificial intelligence to detect hand gestures in defence
Authors: Varsha Naik; Abhishek Chebolu; Janhavi Chavan; Prajakta Chaudhari; Snehalraj Chugh; Ahbaz Memon
Addresses: Dr. Vishwanath Karad MIT World Peace University, Survey No. 124, Paud Rd., Kothrud, Pune, 411038, Maharashtra, India ' Dr. Vishwanath Karad MIT World Peace University, Survey No. 124, Paud Rd., Kothrud, Pune, 411038, Maharashtra, India ' Dr. Vishwanath Karad MIT World Peace University, Survey No. 124, Paud Rd., Kothrud, Pune, 411038, Maharashtra, India ' Dr. Vishwanath Karad MIT World Peace University, Survey No. 124, Paud Rd., Kothrud, Pune, 411038, Maharashtra, India ' Dr. Vishwanath Karad MIT World Peace University, Survey No. 124, Paud Rd., Kothrud, Pune, 411038, Maharashtra, India ' Dr. Vishwanath Karad MIT World Peace University, Survey No. 124, Paud Rd., Kothrud, Pune, 411038, Maharashtra, India
Abstract: Artificial intelligence innovations hold promise for improving military operations while gestures have been the simplest and most powerful medium for communications. This idea proposes a system identifying hand movements and translating them to spoken words, where soldiers on battlegrounds may readily converse with one another. We utilise computational vision, Haar cascade classifiers, CNN, MediaPipe and other approaches for thought patterns through results. There are three techniques in this process: firstly, hand identification system that provides borders around the hand placed to another screen fixing image magnification using OpenCV and Matplot; following with a hand-held skeleton-projected connection model using MediaPipe's mapping system libraries in real-time capturing at 30 fps; lastly, sound element being introduced in each class improving recognition of the gestures. Dataset of around 10,000 images across five different classes are built with CNN architecture used to classify. This research shows results to discover AI options for military applications.
Keywords: military; gesture recognition; object classification; object tracking; Haar cascade; MediaPipe; artificial neural network; convolutional neural network; ConvNet.
DOI: 10.1504/IJCISTUDIES.2022.126906
International Journal of Computational Intelligence Studies, 2022 Vol.11 No.2, pp.94 - 112
Received: 29 Aug 2021
Accepted: 31 Aug 2021
Published online: 11 Nov 2022 *