Human action recognition from video sequences

This PhD research has proposed new machine learning techniques to improve human action recognition based on local features. Several novel video representation and classification techniques have been proposed to increase the performance with lower computational complexity. The major contributions are the construction of new feature representation techniques, based on advanced machine learning techniques such as multiple instance dictionary learning, Latent Dirichlet Allocation (LDA) and Sparse coding. A Binary-tree based classification technique was also proposed to deal with large amounts of action categories. These techniques are not only improving the classification accuracy with constrained computational resources but are also robust to challenging environmental conditions. These developed techniques can be easily extended to a wide range of video applications to provide near real-time performance.

Elliptical summary randomisation for sensor-based human activity recognition

Many conventional statistical machine learning al- gorithms generalise poorly if distribution bias ex- ists in the datasets. For example, distribution bias arises in the context of domain generalisation, where knowledge acquired from multiple source domains need to be used in a previously unseen target domains. We propose Elliptical Summary Randomisation (ESRand), an efficient domain generalisation approach that comprises of a randomised kernel and elliptical data summarisation. ESRand learns a domain interdependent projection to a la- tent subspace that minimises the existing biases to the data while maintaining the functional relationship between domains. In the latent subspace, ellipsoidal summaries replace the samples to enhance the generalisation by further removing bias and noise in the data. Moreover, the summarisation enables large-scale data processing by significantly reducing the size of the data. Through comprehensive analysis, we show that our subspace-based approach outperforms state-of-the-art results on several activity recognition benchmark datasets, while keeping the computational complexity significantly low.

What kind of sick person would see a Foot Doctor! The Foot Disease in Inpatients Study (FDIS)

Background Many different guidelines recommend people with foot complications, or those at risk, should attend multiple health professionals for foot care each year. However, few studies have investigated the characteristics of those attending health professionals for foot care and if those characteristics match those requiring foot care as per guideline recommendations. The aim of this paper was to determine the associated characteristics of people who attended a health professional for foot care in the year prior to their hospitalisation. Methods Eligible participants were all adults admitted overnight, for any reason, into five diverse hospitals on one day; excluding maternity, mental health and cognitively impaired patients. Participants underwent a foot examination to clinically diagnose different foot complications; including wounds, infections, deformity, peripheral arterial disease and peripheral neuropathy. They were also surveyed on social determinant, medical history, self-care, foot complication history, and, past health professional attendance for foot care in the year prior to hospitalisation. Results Overall, 733 participants consented; mean(±SD) age 62(±19) years, 408 (55.8%) male, 172 (23.5%) diabetes. Two hundred and fifty-six (34.9% (95% CI) (31.6-38.4)) participants had attended a health professional for foot care; including attending podiatrists 180 (24.5%), GPs 93 (24.6%), and surgeons 36 (4.9%). In backwards stepwise multivariate analyses attending any health professional for foot care was independently associated (OR (95% CI)) with diabetes (3.0 (2.1-4.5)), arthritis (1.8 (1.3-2.6)), mobility impairment (2.0 (1.4-2.9)) and previous foot ulcer (5.4 (2.9-10.0)). Attending a podiatrist was independently associated with female gender (2.6 (1.7-3.9)), increasing years of age (1.06 (1.04-1.08), diabetes (5.0 (3.2-7.9)), arthritis (2.0 (1.3-3.0)), hypertension (1.7 (1.1-2.6) and previous foot ulcer (4.5 (2.4-8.1). While attending a GP was independently associated with having a foot ulcer (10.4 (5.6-19.2). Conclusions Promisingly these findings indicate that people with a diagnosis of diabetes and arthritis are more likely to attend health professionals for foot care. However, it also appears those with active foot complications, or significant risk factors, may not be more likely to receive the multi-disciplinary foot care recommended by guidelines. More concerted efforts are required to ensure all people with foot complications are receiving recommended foot care.