Engineering Science - Undergraduate Honours Theses

Receive updates for this collection

Preliminary MEG Study of Emotional Face Processing Among Children

Peer reviewed: 
No, item is not peer reviewed.
Date created: 
2021-08-22
Abstract: 

Human social interaction heavily depends on understanding faces, as they carry information about age, gender and emotions. Extensive research has been performed to understand face processing in adults and in children using functional magnetic resonance imaging (fMRI) and electroencephalography (EEG). However, there are far fewer studies using magnetoencephalography (MEG) and only a small handful that study children. The data from this study was one of the first that aimed to find the neurological pathways of face processing of children in the MEG space and additionally aimed to examine the differences in emotional facial presentations. We used MEG datasets from 10 healthy children in the age range of 9 to 16 years old collected during the presentation of static faces with different expressions (neutral, anger and fear) and objects (butterfly, fish and guitar) as control states. The data was preprocessed and an event-related beamformer was utilized to localize the millisecond timescale of activity of the brain processes during the face stimuli and during the object stimuli. We observed activity consistent with the face sensitive M170 event-related  field component with a group average amplitude of 10nA peak and 171 ms post stimulus.  This component localized to the occipital-temporal-parietal region in the right hemisphere consistent with either the fusiform face area (FFA) or the posterior superior temporal sulcus (pSTS).  To our knowledge, this may be one of the few studies that demonstrate localization of face sensitive areas in children using an event-related beamformer approach for MEG. 

Document type: 
Thesis

A Novel Exoskeleton Prototype Based on the Use of IMUs to Track and Mimic Motion

Peer reviewed: 
No, item is not peer reviewed.
Date created: 
2021-07-18
Abstract: 

This thesis presents the development of a person-portable exoskeleton prototype which is designed to be controlled with Inertial Measurement Units (IMUs). It utilizes Euler angles calculated by the IMUs to track the rotation of the user’s forearm and then performs the same rotation, mimicking the user. Special care is taken with the prototype’s control algorithm to ignore changes in Euler angles caused by non-forearm rotations, which can otherwise cause erroneous prototype movements. The prototype is successful in demonstrating this method of control but does require the user to follow some specific guidelines to work at maximum effectiveness. Future iterations of the prototype can be easily improved by replacing some of the commercially available materials with more specialized ad-hoc products.

Document type: 
Thesis

Deep Learning for Satellite Image Analysis

Author: 
Peer reviewed: 
No, item is not peer reviewed.
Date created: 
2017-12-18
Abstract: 

Deep learning architectures have the potential of saving the world from losing football fieldsized forest areas each second. These architectures possess large learning capacities when compared to conventional machine learning architectures, and thus are trained on sizable data-sets to efficiently extract both coarse and fine features from various image scenes. As a result, they can provide crucial information that is needed to manage the deforestation process and its consequences on the environment and ecosystem more effectively. This thesis outlines the two deep learning based systems designed for satellite image analysis. The first system analyzed satellite images of the Amazon, and the goal was to interpret the image content by providing a set of labels that best describe it. The highest performing architecture was able to achieve a score of 92.886% while a combination of several high performance, yet uncorrelated, architectures increased the overall score to 93.070%. This result is only 0.248% lower than what current state of the art algorithms achieved on the same task. The second system was designed to detect the presence of clouds in Landsat 8 images by analyzing small chips within each large image. This system produced cloud masks, which were then compared to the corresponding ground truth cloud masks obtained from the provided images. The predicted cloud masks were able to achieve an average score of 92.931%, which is very high for the given accuracy measure.

Document type: 
Thesis
File(s): 
Supervisor(s): 
Ivan V. Bajić
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
Honours Bachelor of Applied Science

Implementation of Active Noise Cancellation in a Duct

Author: 
Peer reviewed: 
No, item is not peer reviewed.
Date created: 
2013-11-15
Abstract: 

An Active Noise Cancellation (ANC) system is implemented in real time using both feed forward LFXLMS (Leaky filtered-X least-mean-square) and feedback LFXLMS approaches for adaptive filtering. ANC algorithms are implemented on a ADAU1446 evaluation board and tested in terms of sound cancellation in a duct. The hardware and software interfaces required for the system are explained in detail. A test bed is developed to measure the performance of sound cancellation. Results are analysed in detail and recommendations are made for future research work to improve the performance of the system and to realize noise cancellation in 3D space.

Document type: 
Thesis
File(s):