The project concerns the assessment of foot disease using load measurement systems which obtains 3 axis loads information across an array of measurement points.
Both systems will produce load maps – in slightly different shapes but essentially the same; N nodes each measuring load (force) in 3 axes XYZ. To develop this research we currently have a 64 node system “MagBoard” and two simulator systems which mimic footsteps on the sensing surface.
The main focus of your project is to apply machine learning techniques (e.g. neural networks) to analyse the ‘load maps’ obtained from these sensory systems. In particular, we’re interested in being able to:
Differentiate (and classify) healthy and diseased foot profiles (we will make model feet with different mechanical properties / shapes to represent this)
In this research, we have proposed a Bidirectional LSTM (Long Short Term Memory) network for this binary classification. LSTM network is most suitable model for this dataset as it allows sequence data as an input, and make prediction based on the individual time steps of sequence data.
I would like to thank Dr. Peter Culmer, University of Leeds for guiding me throughout my internship at SurgTech Lab, University of Leeds.