Foreword:
The purpose of this session is twofold. Firstly, it aims to summarize the work of COMP6620 students over the past semester in order to provide an anchor. Secondly, it aims to sensitize the wider public that useful work exists which could immediately resolve some existing challenges, either by refining outcomes to be presented, or by exploiting available capabilities.
The course COMP6620:Computer Vision covers some basic image processing concepts and applications, while recognizing that one semester is too short for an indepth study of all the relevant theories and ideas. As such, a bloc of time in the course is reserved for each student to perform guided exploration in state of the art techniques and applications, in order to present a customized solution.
These presentations are the fruits of that process. Extended abstractions are also available in PDF. It is possible that, for some of these projects, this presentation could signal the end. It is also possible that work here could represent seeds planted that blossom into new applications, or expose subproblems that could give rise to ground breaking pure research. So, enjoy the sessions below, and feel free to reach out to connect! Paul Gaynor(PhD), Lecturer, COMP6620:Computer Vision.
Presentations:
-
Shooting Star Catcher: An Efficient Neural Network Architecture to Detect Meteors using an All Sky camera.
| Jevaughn Dixon
-
Adaptive Cross-Domain Dehazing (ACD): A Multi-Domain Approach to Image Dehazing.
| Golda-Mae Bullock
-
No Fake Patois: Visualized Voice Profiles for Accent Verification
| Ackeem Lawrence
-
Ensemble Convolutional Neural Networks (CNNs) for Enhanced Diagnosis of Chest Diseases from Medical Images
| Tamara Matthews
-
ECAS:An Embedded Cornering Assistance system to Decrease the Likelihood of Fatal Motorcycle Accidents.
| Jesse Roye
-
SeeU_HealU :Exploiting Human Pose Feedback to Augment Physiotherapy
| Danielle Gilman
-
Classification of Football Tactics using Video Analysis
| Rushane Barnes
-
Never2Far2Fix:A finetuned CNN to identify spacecraft in need of repair
| Ashley Wright
-
Detecting ADS-B Spoofing with A Self- Conditioning Transformer Model
| Navion Plummer
-
Seeing through the Past: Autoencoder based architecture to reconstruct artefacts from degraded fragments
| Chad Brown
-
Tracking Football Players from Multi—Media sources.
| Daniel Haye
-
AssuredDriver: Improving vehicle safety by leveraging Difussion Models on environmental images
| Chevaughn Johnson
-
Injury Detection using Pose Estimation and Detection
| Damion Lowers
-
Form Reader: Combining Optical Character Recognition (OCR) and
Text-to-speech (TTS)Technology to increase accessibility to
information for those with learning challenges
| Kevaughn Tyrell
-
Leveraging TensorFlow for Efficient Brain Tumor Classification in MRI Images
| Zinedine Lloyd
-
Sentiment Analysis on Extracted Text using SVM and Transfer Learning
| Javid Bell
-
Goal!: Efficient Video Analysis to Detect when a Football Player Shoots at Goal
| Matthew Levy
