A5034501904- Emotion and Mood Recognition
- Gaze Tracking and Assistive Technology
- Face recognition and analysis
- Digital Media Forensic Detection
- Generative Adversarial Networks and Image Synthesis
- Visual Attention and Saliency Detection
- Retinal Imaging and Analysis
- Face Recognition and Perception
- Speech and Audio Processing
- Mental Health Research Topics
- Glaucoma and retinal disorders
- Advanced Computing and Algorithms
- Color perception and design
- IoT and Edge/Fog Computing
- Video Surveillance and Tracking Methods
- EEG and Brain-Computer Interfaces
- Indoor and Outdoor Localization Technologies
- Face and Expression Recognition
- Advanced Image Processing Techniques
- Human Pose and Action Recognition
- Sleep and Work-Related Fatigue
- Anomaly Detection Techniques and Applications
- Pain Mechanisms and Treatments
- Hand Gesture Recognition Systems
- Musculoskeletal pain and rehabilitation
Curtin University
2023-2025
Australian Regenerative Medicine Institute
2021-2023
Monash University
2020-2023
Pennsylvania State University
2022
Indian Institute of Technology Ropar
2017-2021
Purdue University West Lafayette
2020
University of Potsdam
2020
Research in automatic affect recognition has come a long way. This paper describes the fifth Emotion Recognition Wild (EmotiW) challenge 2017. EmotiW aims at providing common benchmarking platform for researchers working on different aspects of affective computing. year there are two sub-challenges: a) Audio-video emotion and b) group-level recognition. These challenges based acted facial expressions wild group databases, respectively. The particular focus is to evaluate method `in wild'...
Depression has become a big problem in our society today. It is also major reason for suicide, especially among teenagers. In the current outbreak of coronavirus disease (COVID-19), affected countries have recommended social distancing and lockdown measures. Resulting interpersonal isolation, these measures raised serious concerns mental health depression. Generally, clinical psychologists diagnose depressed people via face-to-face interviews following depression criteria. However, often...
This paper proposes a self-supervised approach to learn universal facial representations from videos, that can transfer across variety of analysis tasks such as Facial Attribute Recognition (FAR), Expression (FER), DeepFake Detection (DFD), and Lip Synchronization (LS). Our proposed framework, named MARLIN, is video masked autoencoder, learns highly robust generic embeddings abundantly available non-annotated web crawled videos. As challenging auxiliary task, MARLIN reconstructs the...
Most deepfake detection methods focus on detecting spatial and/or spatio-temporal changes in facial attributes and are centered around the binary classification task of whether a video is real or fake. This because available benchmark datasets contain mostly visual-only modifications present entirety video. However, sophisticated may include small segments audio audio-visual manipulations that can completely change meaning content. To addresses this gap, we propose new dataset, Localized...
Robust gaze estimation is a challenging task, even for deep CNNs, due to the non-availability of large-scale labeled data. Moreover, annotation time-consuming process and requires specialized hardware setups. We propose MTGLS: Multi-Task Gaze framework with Limited Supervision, which leverages abundantly available non-annotated facial image MTGLS distills knowledge from off-the-shelf analysis models, learns strong feature representations human eyes, guided by three complementary auxiliary...
This study investigates the effects of cognitive bias awareness, feedback and adaptation, career confidence, academic expectations, habits, self-perceived knowledge on decision-making risk-taking behaviors students at a higher educational level. For study’s purposes, data were collected through structured questionnaire from sample 435 respondents tertiary Bangladesh, who chosen stratified random technique to produce diverse representation. Then, demographic analyses conducted, multiple...
This paper proposes a pipeline for automatic group-level affect analysis. A deep neural network-based approach, which leverages on the facial-expression information, scene information and high-level facial visual attribute is proposed. capsule architecture used to predict expression. Transfer learning Inception-V3 extract global image-based features contain information. Another network trained inferring attributes of group members. Further, these are pooled at train affect. The prediction...
Labelling of human behavior analysis data is a complex and time consuming task. In this paper, fully automatic technique for labelling an image based gaze dataset driver zone estimation proposed. Domain knowledge added to the recording paradigm later labels are generated in manner using Speech To Text conversion (STT). order remove noise STT process due different illumination ethnicity subjects our data, speech frequency energy analysed. The resultant Driver Gaze Wild (DGW) contains 586...
This paper proposes a database for group level emotion recognition in videos. The motivation is coming from the large number of information which users are sharing online. gives us opportunity to use this perceived affect various tasks. Most work area has been restricted controlled environments. In paper, we explore and cohesion real-world environment. There several challenges involved moving environment scenarios such as face tracking limitations, illumination variations, occlusion type...
The cohesiveness of a group is an essential indicator the emotional state, structure and success people. We study factors that influence perception group-level cohesion propose methods for estimating human-perceived on scale. In order to identify visual cues (attributes) cohesion, we conducted user survey. Image analysis performed at via multi-task convolutional neural network. For analyzing contribution facial expressions members predicting Group Cohesion Score (GCS), capsule network...
Automatic eye gaze estimation has interested researchers for a while now. In this paper, we propose an unsupervised learning based method estimating the region. To train proposed network "Ize-Net" in self-supervised manner, collect large `in wild' dataset containing 1,54,251 images from web. For database, divide into three regions on automatic technique pupil-centers localization and then use feature-based to determine The performance is evaluated Tablet Gaze CAVE datasets by fine-tuning...
Eye gaze analysis is an important research problem in the field of Computer Vision and Human-Computer Interaction. Even with notable progress last 10 years, automatic still remains challenging due to uniqueness eye appearance, eye-head interplay, occlusion, image quality, illumination conditions. There are several open questions, including what cues interpret direction unconstrained environment without prior knowledge how encode them real-time. We review across a range tasks applications...
The detection and localization of deepfake content, particularly when small fake segments are seamlessly mixed with real videos, remains a significant challenge in the field digital media security. Based on recently released AV-Deepfake1M dataset, which contains more than 1 million manipulated videos across 2,000 subjects, we introduce 1M-Deepfakes Detection Challenge. This is designed to engage research community developing advanced methods for detecting localizing manipulations within...
This article discusses the prediction of cohesiveness a group people in images. The is an essential indicator emotional state, structure, and success group. We study factors that influence perception group-level cohesion propose methods for estimating human-perceived on scale. To identify visual cues (attributes) cohesion, we conducted user survey. Image analysis performed at via multi-task convolutional neural network. A capsule network explored analyzing contribution facial expressions...
Empathy is a social skill that indicates an individual's ability to understand others. Over the past few years, empathy has drawn attention from various disciplines, including but not limited Affective Computing, Cognitive Science and Psychology. context-dependent term; thus, detecting or recognising potential applications in society, healthcare education. Despite being broad overlapping topic, avenue of detection studies leveraging Machine Learning remains underexplored holistic literature...
The detection and localization of highly realistic deepfake audio-visual content are challenging even for the most advanced state-of-the-art methods. While research efforts in this domain focused on detecting high-quality images videos, only a few works address problem small segments manipulations embedded real videos. In research, we emulate process such generation propose AV-Deepfake1M dataset. dataset contains content-driven (i) video manipulations, (ii) audio (iii) more than 2K subjects...
Deep Neural Networks (DNNs) achieve state-of-the-art accuracy in many computer vision tasks, such as object counting. Object counting takes two inputs: an image and query reports the number of occurrences queried object. To high accuracy, DNNs require billions operations, making them difficult to deploy on resource-constrained, low-power devices. Prior work shows that a significant DNN operations are redundant can be eliminated without affecting accuracy. reduce these redundancies, we...