Chris McCarthy
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Chris McCarthy has a total of 100 co-authors. Here are the top co-authors by number of shared publications:
- Felip Martí
Swinburne Univ Technol, John St, Hawthorn, Vic 3122, Australia
4 shared publications - Lau Bee Theng
Faculty of Engineering, Computing and Science, Swinburne University of Technology, Kuching, Sarawak, Malaysia
4 shared publications - Dana Rezazadegan
Swinburne University of Technology
3 shared publications - Mark Tee Kit Tsun
Faculty of Engineering, Computing and Science, Swinburne University of Technology Sarawak Campus, Kuching 93350, Malaysia
3 shared publications - Abhik Banerjee
SoCET, Swinburne University of Technology Faculty of Science Engineering and Technology
2 shared publications - Hadi Ghaderi
School of Business, Law and Entrepreneurship, Swinburne University of Technology, Hawthorn, VIC, Australia
2 shared publications
Latest Publications
- Open Access
- 16 January 2026
- Published by Elsevier in Electric Power Systems Research
- Vol. 255
Electricity price spike forecasting via multivariate adaptive visibility graph neural networks
- 16 January 2026
- Published by Elsevier in Electric Power Systems Research
- Vol. 255
Open AccessElectricity price spike forecasting via multivariate adaptive visibility graph neural networks
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 150536-150556
Reflective-Surface Failure Verifiability for Indoor LiDAR-Vision-IMU Mobile Robot Navigation
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 150536-150556
Open AccessReflective-Surface Failure Verifiability for Indoor LiDAR-Vision-IMU Mobile Robot Navigation
- Open Access
- 2 December 2025
- Published by Elsevier in Forensic Science International
- Vol. 379, 112760
Forensic gender and stature identification from footprint images using machine learning
- 2 December 2025
- Published by Elsevier in Forensic Science International
- Vol. 379, 112760
Open AccessForensic gender and stature identification from footprint images using machine learning
- Open Access
- 1 August 2025
- Published by Springer Nature in Neural Computing and Applications
- Vol. 37 (26) , 21969-21995
An evaluation of a pre-trained transformer-based self-distillation model (DINOv2) for cross-domain plant species identification
Plant species identification is a fundamental process in botany and agriculture sector. In recent years, deep neural networks have become the primary approach for automating this task, providing valuable insights into biodiversity, ecological systems, and agricultural practices. Along with more discoveries in plant species, training a deep neural network becomes very challenging as the cost required to collect and annotate plant samples is expensive and impractical. Despite the lack of labelled plant samples, recent studies have explored the potential of leveraging publicly available and systematically annotated plant specimens in herbaria coupled with field images for plant species identification through cross-domain adaptation techniques. However, the accuracy of these methods remains unsatisfactory, motivating the exploration of alternative approaches. In this paper, we evaluated the feasibility of employing a pre-trained transformer-based self-distillation model (DINOv2) for cross-domain plant species identification tasks. We trained our model with the PlantCLEF2020 dataset comprised of approximately 320 k herbarium and field images representing 997 plant species. Our approach leverages the advanced feature extraction capabilities of DINOv2, which enhances cross-domain adaptation by effectively bridging the gap between herbarium and field images, achieving a 17.7% improvement over the best model proposed in previous work, that employs ensembles of Siamese network architectures with triplet loss (HFTL-ENS and OSM-ENS).
- 1 August 2025
- Published by Springer Nature in Neural Computing and Applications
- Vol. 37 (26) , 21969-21995
Open AccessAn evaluation of a pre-trained transformer-based self-distillation model (DINOv2) for cross-domain plant species identification
Plant species identification is a fundamental process in botany and agriculture sector. In recent years, deep neural networks have become the primary approach for automating this task, providing valuable insights into biodiversity, ecological systems, and agricultural practices. Along with more discoveries in plant species, training a deep neural network becomes very challenging as the cost required to collect and annotate plant samples is expensive and impractical. Despite the lack of labelled plant samples, recent studies have explored the potential of leveraging publicly available and systematically annotated plant specimens in herbaria coupled with field images for plant species identification through cross-domain adaptation techniques. However, the accuracy of these methods remains unsatisfactory, motivating the exploration of alternative approaches. In this paper, we evaluated the feasibility of employing a pre-trained transformer-based self-distillation model (DINOv2) for cross-domain plant species identification tasks. We trained our model with the PlantCLEF2020 dataset comprised of approximately 320 k herbarium and field images representing 997 plant species. Our approach leverages the advanced feature extraction capabilities of DINOv2, which enhances cross-domain adaptation by effectively bridging the gap between herbarium and field images, achieving a 17.7% improvement over the best model proposed in previous work, that employs ensembles of Siamese network architectures with triplet loss (HFTL-ENS and OSM-ENS).
- Open Access
- 1 July 2025
- Published by SAGE Publications in Journal of Visual Impairment & Blindness
- Vol. 119 (4) , 245-255
Assessment of Low Vision Rehabilitation Goals Using the Canadian Occupational Performance Measure
Introduction: As emerging therapies enter clinical trials, functional vision assessments are required by regulatory bodies to show real-world effectiveness. Development of new outcome measures should begin with a thorough understanding of the goals of people with low vision. The study aims to explore the rehabilitation goals of people with low vision by analyzing data collected with the Canadian Occupational Performance Measure (COPM) in an Australian cohort. Methods: We applied thematic analysis to a retrospective dataset of reported goals of people with low vision, obtained using the COPM. The dataset was obtained from a low vision service provider, Vision Australia. We analyzed de-identified COPM data from 190 clients (122 female, 64%). Three main themes were identified (safety and confidence, life skills, and leisure). Results: “Safety and confidence” is about participants being able to access their home, community, and workplace safely and confidently. Most (87%) of the goals related to effective navigation and mobility. The “life skills” theme included goals related to personal care (25%), kitchen skills (25%), and interpersonal skills (5%), and for parents, their child's development (43%). The “leisure” theme is about engaging in leisure activities and socializing with others. More than half (53%) of the participants reported access to media as a rehabilitation goal, and many wanted to improve their information technology skills. Discussion: In conclusion, we identified three main goals: safety and confidence, especially in accessing the wider community and safety around the home environment; life skills, including kitchen work; and leisure activity, such as accessing print and electronic material. Most mobility goals focused on enhancing feelings of safety, confidence, and independence. These findings underscore the crucial role of emotional support in enhancing functional mobility. Implications for Practitioners: The results identify important goals that could be incorporated into vision restoration clinical trials, in order to emphasize clients’ desired visual functioning.
- 1 July 2025
- Published by SAGE Publications in Journal of Visual Impairment & Blindness
- Vol. 119 (4) , 245-255
Open AccessAssessment of Low Vision Rehabilitation Goals Using the Canadian Occupational Performance Measure
Introduction: As emerging therapies enter clinical trials, functional vision assessments are required by regulatory bodies to show real-world effectiveness. Development of new outcome measures should begin with a thorough understanding of the goals of people with low vision. The study aims to explore the rehabilitation goals of people with low vision by analyzing data collected with the Canadian Occupational Performance Measure (COPM) in an Australian cohort. Methods: We applied thematic analysis to a retrospective dataset of reported goals of people with low vision, obtained using the COPM. The dataset was obtained from a low vision service provider, Vision Australia. We analyzed de-identified COPM data from 190 clients (122 female, 64%). Three main themes were identified (safety and confidence, life skills, and leisure). Results: “Safety and confidence” is about participants being able to access their home, community, and workplace safely and confidently. Most (87%) of the goals related to effective navigation and mobility. The “life skills” theme included goals related to personal care (25%), kitchen skills (25%), and interpersonal skills (5%), and for parents, their child's development (43%). The “leisure” theme is about engaging in leisure activities and socializing with others. More than half (53%) of the participants reported access to media as a rehabilitation goal, and many wanted to improve their information technology skills. Discussion: In conclusion, we identified three main goals: safety and confidence, especially in accessing the wider community and safety around the home environment; life skills, including kitchen work; and leisure activity, such as accessing print and electronic material. Most mobility goals focused on enhancing feelings of safety, confidence, and independence. These findings underscore the crucial role of emotional support in enhancing functional mobility. Implications for Practitioners: The results identify important goals that could be incorporated into vision restoration clinical trials, in order to emphasize clients’ desired visual functioning.
- 15 June 2025
- Published by Springer Nature
- p. 278-290
Linked Adapters: Linking Past and Future to Present for Effective Continual Learning
- 15 June 2025
- Published by Springer Nature
- p. 278-290
Linked Adapters: Linking Past and Future to Present for Effective Continual Learning
- Open Access
A Lightweight Large Vision-language Model for Multimodal Medical Images
Medical Visual Question Answering (VQA) enhances clinical decision-making by enabling systems to interpret medical images and answer clinical queries. However, developing efficient, high-performance VQA models is challenging due to the complexity of medical imagery and diverse modalities. In this paper, we introduce a lightweight, multimodal VQA model integrating BiomedCLIP for image feature extraction and LLaMA-3 for text processing. Designed for medical VQA tasks, our model achieves state-of-the-art performance on the OmniMedVQA dataset. With approximately 8 billion parameters, it requires only two NVIDIA 40 GB A100 GPUs, demonstrating superior efficiency over larger models. Our results show 73.4% accuracy for open-end questions, surpassing existing models and validating its potential for real-world medical applications. Key contributions include a specialized multimodal VQA model, a resource-efficient architecture, and strong performance in answering open-ended clinical questions.
Open AccessA Lightweight Large Vision-language Model for Multimodal Medical Images
Medical Visual Question Answering (VQA) enhances clinical decision-making by enabling systems to interpret medical images and answer clinical queries. However, developing efficient, high-performance VQA models is challenging due to the complexity of medical imagery and diverse modalities. In this paper, we introduce a lightweight, multimodal VQA model integrating BiomedCLIP for image feature extraction and LLaMA-3 for text processing. Designed for medical VQA tasks, our model achieves state-of-the-art performance on the OmniMedVQA dataset. With approximately 8 billion parameters, it requires only two NVIDIA 40 GB A100 GPUs, demonstrating superior efficiency over larger models. Our results show 73.4% accuracy for open-end questions, surpassing existing models and validating its potential for real-world medical applications. Key contributions include a specialized multimodal VQA model, a resource-efficient architecture, and strong performance in answering open-ended clinical questions.
- Open Access
- 1 January 2025
- Published by HICSS Conference Office
Improving Vision-Based Freight Vehicle Detection in Smart Cities Using Generative Adversarial Networks
A freight activity grows within urban areas, accurately estimating their impact becomes crucial for effective planning and modelling of transport infrastructure. Reliable Origin-Destination (OD) information is essential for strategic transport models, which guide future infrastructure investments and sustainable urban development. In this paper we propose a Generative Adversarial Network (GAN)-based domain adaptation approach for accurately detecting heavy vehicles from low-quality surveillance data under varying lighting conditions. To the best of our knowledge, this is the first study to use GAN-based data augmentation for freight vehicle movement analysis in smart cities, contributing to the development of more efficient, scalable, and reliable smart city planning solutions.
- 1 January 2025
- Published by HICSS Conference Office
Open AccessImproving Vision-Based Freight Vehicle Detection in Smart Cities Using Generative Adversarial Networks
A freight activity grows within urban areas, accurately estimating their impact becomes crucial for effective planning and modelling of transport infrastructure. Reliable Origin-Destination (OD) information is essential for strategic transport models, which guide future infrastructure investments and sustainable urban development. In this paper we propose a Generative Adversarial Network (GAN)-based domain adaptation approach for accurately detecting heavy vehicles from low-quality surveillance data under varying lighting conditions. To the best of our knowledge, this is the first study to use GAN-based data augmentation for freight vehicle movement analysis in smart cities, contributing to the development of more efficient, scalable, and reliable smart city planning solutions.
- Open Access
Linked Adapters: Linking Past and Future to Present for Effective Continual Learning
Continual learning allows the system to learn and adapt to new tasks while retaining the knowledge acquired from previous tasks. However, deep learning models suffer from catastrophic forgetting of knowledge learned from earlier tasks while learning a new task. Moreover, retraining large models like transformers from scratch for every new task is costly. An effective approach to address continual learning is to use a large pre-trained model with task-specific adapters to adapt to the new tasks. Though this approach can mitigate catastrophic forgetting, they fail to transfer knowledge across tasks as each task is learning adapters separately. To address this, we propose a novel approach Linked Adapters that allows knowledge transfer through a weighted attention mechanism to other task-specific adapters. Linked adapters use a multi-layer perceptron (MLP) to model the attention weights, which overcomes the challenge of backward knowledge transfer in continual learning in addition to modeling the forward knowledge transfer. During inference, our proposed approach effectively leverages knowledge transfer through MLP-based attention weights across all the lateral task adapters. Through numerous experiments conducted on diverse image classification datasets, we effectively demonstrated the improvement in performance on the continual learning tasks using Linked Adapters.
Open AccessLinked Adapters: Linking Past and Future to Present for Effective Continual Learning
Continual learning allows the system to learn and adapt to new tasks while retaining the knowledge acquired from previous tasks. However, deep learning models suffer from catastrophic forgetting of knowledge learned from earlier tasks while learning a new task. Moreover, retraining large models like transformers from scratch for every new task is costly. An effective approach to address continual learning is to use a large pre-trained model with task-specific adapters to adapt to the new tasks. Though this approach can mitigate catastrophic forgetting, they fail to transfer knowledge across tasks as each task is learning adapters separately. To address this, we propose a novel approach Linked Adapters that allows knowledge transfer through a weighted attention mechanism to other task-specific adapters. Linked adapters use a multi-layer perceptron (MLP) to model the attention weights, which overcomes the challenge of backward knowledge transfer in continual learning in addition to modeling the forward knowledge transfer. During inference, our proposed approach effectively leverages knowledge transfer through MLP-based attention weights across all the lateral task adapters. Through numerous experiments conducted on diverse image classification datasets, we effectively demonstrated the improvement in performance on the continual learning tasks using Linked Adapters.
- Open Access
- 9 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. PP, 1
A Multi-Stream Approach to Mixed-Traffic Accident Recognition Using Deep Learning
Road traffic accidents are a leading cause of injuries and fatalities globally, prompting extensive research into deep learning-based accident recognition models for their superior performance in computer vision tasks. However, most studies focus on non-mixed traffic environments, where detection is simpler due to predictable traffic patterns and uniform vehicle types. In contrast, mixed-traffic scenarios present greater challenges as diverse vehicles, motorcyclists, and pedestrians move unpredictably. Models relying on a single type of perception are effective in structured traffic but struggle to handle the complexities of mixed-traffic environments. This study proposes a novel multi-stream deep learning model called Accident Recognition in Mixed-Traffic Scene (ARMS), which integrates three distinct streams: the first stream analyzes the overall accident scene, the second focuses on mixed-traffic accident features, and the third examines vehicle motion abnormalities through object detection and tracking, aimed at improving road accident recognition accuracy in mixed-traffic environments at intersections. This model is trained and evaluated using datasets from CADP, UA-DETRAC, and supplementary online sources. The results demonstrate that the ARMS model achieves an accuracy of 93.3%, with performance improving significantly through the fusion of the individual streams. Additionally, the ARMS model was evaluated using two publicly available standard datasets, which further highlights its improved performance in recognizing mixed-traffic accidents compared to existing studies.
- 9 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. PP, 1
Open AccessA Multi-Stream Approach to Mixed-Traffic Accident Recognition Using Deep Learning
Road traffic accidents are a leading cause of injuries and fatalities globally, prompting extensive research into deep learning-based accident recognition models for their superior performance in computer vision tasks. However, most studies focus on non-mixed traffic environments, where detection is simpler due to predictable traffic patterns and uniform vehicle types. In contrast, mixed-traffic scenarios present greater challenges as diverse vehicles, motorcyclists, and pedestrians move unpredictably. Models relying on a single type of perception are effective in structured traffic but struggle to handle the complexities of mixed-traffic environments. This study proposes a novel multi-stream deep learning model called Accident Recognition in Mixed-Traffic Scene (ARMS), which integrates three distinct streams: the first stream analyzes the overall accident scene, the second focuses on mixed-traffic accident features, and the third examines vehicle motion abnormalities through object detection and tracking, aimed at improving road accident recognition accuracy in mixed-traffic environments at intersections. This model is trained and evaluated using datasets from CADP, UA-DETRAC, and supplementary online sources. The results demonstrate that the ARMS model achieves an accuracy of 93.3%, with performance improving significantly through the fusion of the individual streams. Additionally, the ARMS model was evaluated using two publicly available standard datasets, which further highlights its improved performance in recognizing mixed-traffic accidents compared to existing studies.