For my research, I had to acquire multiple dataset to analyse and investigate the developed algorithms.
A large multimodal dataset for remote healthcare monitoring of people living with dementia, including daily activity, sleep, physiological signals, and clinical events collected in-home over multiple weeks.
A multi-session sEMG dataset designed to evaluate the repeatability of hand movement classification for prosthetic control, featuring data from 10 subjects across 100 acquisitions. DB6 supports research on robustness, cross-session adaptation, and sEMG-based hand gesture recognition.
Localising and recognising the presence of mechanical fractures is an important task necessary in hazardous environments during waste decommissioning. It is especially useful to avoid spillage from containers keeping chemical and radioactive waste or to identify concrete fractures at early stages, to prevent their growth which may lead to larger macro-scale catastrophic failures. This dataset consists of 3000 labelled images of fractures and 24000 labelled augmented images. The acquired data can be used for fracture localisation with object detection.
Tactile and proximity sensing can be efficiently used to perform automatic crack detection. This dataset was acquired with a fibre optic based finger-shaped sensor which slided on a total of 10 different 3D printed surfaces for a total of 50 acquistions. The acquired data can be used for fracture detection on surfaces using both tactile and proximity features.