COMPLEX-VALUED CONVOLUTIONAL NEURAL NETWORK CLASSIFICATION OF HAND GESTURES FROM RADAR IMAGES

Abstract

Hand gesture recognition systems have yielded many exciting advancements in the last decade and become more popular in HCI (human-computer interaction) with several application areas, which spans from safety and security applications to automotive field. Various deep neural network architectures have already been inspected for hand gesture recognition systems, including multi-layer perceptron (MLP) [5], convolutional neural network (CNN) [6], recurrent neural network (RNN) [7] and a cascade of the last two architectures known as CNN-RNN [8]. However, a major problem still exists, which is most of the existing ML algorithms are designed and developed the building blocks and techniques for real- valued (RV). Researchers applied various RV techniques on the complex-valued (CV) radar images, such as converting a CV optimisation problem into a RV one, by splitting the complex numbers into their real and imaginary parts. However, the major disadvantage of this method is that, the resulting algorithm will double the network dimensions. Recent work on RNNs and other fundamental theoretical analysis suggest that, CV numbers have a richer representational capacity, but due to the absence of the building blocks required to design such models, the performance of CV networks are marginalised. In this report, first we review the background of ML and artificial neural networks (ANNs) in chapter two, then in the third chapter, we explain the characteristics of our utilised two sets of CV datasets. In the fourth chapter, we propose a fully CV-CNN, including all building blocks, forward and backward operations, and derivatives all in complex domain. Then we implement the designed model in Python from scratch and fully in complex domain. We explore the proposed classification model on two sets of CV hand gesture radar images in comparison with the equivalent RV model. In chapter five, we propose a CV-forward residual network, for the purpose of binary classification of the two sets of CV hand gesture radar datasets. We demonstrate the blocks and operations that implement the CV simulated calculations, however, the BP(back propagation) calculation is all in RV domain. Then, we explore the proposed classification model on two sets of CV hand gesture radar images in comparison with the equivalent RV residual model. In chapter six, we propose a CV-forward CNN, which implements the simulated CV operations in the building blocks, however the BP operations are all in RV domain. Then, we explore the proposed classification model on two sets of CV hand gesture radar images in comparison with the equivalent RV-CNN model. At the end, we compare and analyse all three proposed models results and recommend future works.

Date of Award	1 Aug 2024
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Fumie Costen (Supervisor) & Farshad Arvin (Supervisor)