Turning simulated and real visual data into actionable insights using deep learning, automated interpretation, and predictive analytics.
The final stage of the pipeline is where intelligence is extracted. Our AI and computer vision research takes the structured data produced upstream and transforms it into decisions — detecting faults, recognising actions, predicting failures, and guiding autonomous systems.
AI and computer vision are the culmination of the INSIGHT pipeline. The systems developed here consume structured data and digital twins to deliver real-world decisions — from defect classification to autonomous navigation.
Models, frameworks and systems we build and deploy
Deep learning models for detecting, localising, and classifying objects in images and video streams — from infrastructure defects to safety hazards on industrial sites.
Skeleton-based and appearance-based action recognition models for monitoring worker behaviour, detecting safety violations, and analysing physical performance in real-time video.
Forecasting models that learn from historical inspection, sensor, and simulation data to predict structural degradation, equipment failure, and maintenance requirements.
Automated identification and segmentation of corrosion, cracking, delamination, and structural defects in inspection imagery from pipelines, offshore platforms, and civil infrastructure.
Deployable deep learning systems that automatically detect and classify defects in pipeline inspection video, satellite imagery, and drone-captured footage.
Deep learning models trained on medical imaging datasets to assist radiologists in detecting anomalies, segmenting organs, and prioritising clinical review queues.
Perception and decision-making systems for autonomous vehicles, robotic platforms, and UAVs — enabling safe navigation, obstacle avoidance, and task execution.
Loading projects…
Loading team…