A5100681207- Multimodal Machine Learning Applications
- Topic Modeling
- Advanced Image and Video Retrieval Techniques
- Face recognition and analysis
- Colorectal Cancer Screening and Detection
- Artificial Intelligence in Healthcare
- Human Pose and Action Recognition
- Generative Adversarial Networks and Image Synthesis
- Machine Learning in Healthcare
- Natural Language Processing Techniques
- Domain Adaptation and Few-Shot Learning
- Infrared Thermography in Medicine
- Advanced Image Processing Techniques
- Dental Research and COVID-19
- Facial Nerve Paralysis Treatment and Research
Curtin University
2025
Australian National University
2020-2022
Commonwealth Scientific and Industrial Research Organisation
2022
Murdoch University
2018-2019
This article investigates the critical issue of dataset bias in medical imaging, with a particular emphasis on racial disparities caused by uneven population distribution collection. Our analysis reveals that segmentation datasets are significantly biased, primarily influenced demographic composition their collection sites. For instance, Scanning Laser Ophthalmoscopy (SLO) fundus collected United States predominantly feature images White individuals, minority groups underrepresented....
Interactive facial image manipulation attempts to edit single and multiple face attributes using a photo-realistic and/or semantic mask as input. In the absence of (only sketch/mask available), previous methods only retrieve original but ignore potential aiding model controllability diversity in translation process. This paper proposes sketch-to-image generation framework called S2FGAN, aiming improve users' ability interpret flexibility attribute editing from simple sketch. First, restore...
In a typical image captioning pipeline, Convolutional Neural Network (CNN) is used as the encoder and Long Short-Term Memory (LSTM) language decoder. LSTM with attention mechanism has shown remarkable performance on sequential data including captioning. can retain long-range dependency of data. However, it hard to parallelize computations because its inherent characteristics. order address this issue, recent works have benefits in using self-attention, which highly parallelizable without...
Generating a description of an image is called captioning. Image captioning requires to recognize the important objects, their attributes and relationships in image. It also needs generate syntactically semantically correct sentences. Deep learning-based techniques are capable handling complexities challenges In this survey paper, we aim present comprehensive review existing deep techniques. We discuss foundation analyze performances, strengths limitations. datasets evaluation metrics...
<div>The smiles stimuli were collected from the UvA-NEMO dataset. A total of 60 videos randomly selected for this experiment. We created four categories out these videos, namely PV (paired videos), PI images), SV (single video), and SI image). When same person was viewed by observers in both real posed smile we report as "paired videos", otherwise use term "single video".</div>
Interactive facial image manipulation attempts to edit single and multiple face attributes using a photo-realistic and/or semantic mask as input. In the absence of (only sketch/mask available), previous methods only retrieve original but ignore potential aiding model controllability diversity in translation process. This paper proposes sketch-to-image generation framework called S2FGAN, aiming improve users' ability interpret flexibility attribute editing from simple sketch. The proposed...