At the heart of the IVCC Group's mission is the development of machine learning models specialised in video super-resolution (VSR), championed by our dedicated PhD students. These models are meticulously designed for high efficiency and minimal data reliance, significantly reducing computational load and data requirements. This approach not only enhances video quality but also ensures adaptability across various bandwidths and devices. Our team's innovative work in this area is setting new standards in video enhancement, striving for a balance between advanced analytics and sustainable resource utilisation, thereby making sophisticated video processing techniques more accessible and practical for widespread use.
Funded by Enterprise Ireland Innovation Partnership Programme for a few years (since 2018). In a recent project, led by Dr Ye, and an engineer, focuses on multi-vehicle tracking in low-frame rate video environments. We emphasise using synthetic datasets to reduce data labelling costs and are developing streamlined predictors as cost-effective alternatives to traditional methods like the Kalman filter. These innovations aim to reduce computational requirements while providing high tracking and counting accuracy, making our solutions ideal for urban planning and traffic management applications where resources are limited.
Led by a dedicated PhD student, this object involves exploring the security aspects of industrial image processing. We are identifying novel ways to safeguard image data against adversarial backdoor threats and vulnerabilities.