Project Abstract

This is a python based project aimed to evaluate the effectiveness of LSTM due to the limitations of Convolutional Neural Networks (CNNs) in analysing spatiotemporal data. by utilizing Long Short-Term Memory (LSTM) networks, which are adept at handling sequential or time-series data. With CNNs predominantly tailored for image data analysis, LSTM networks offer a unique opportunity to delve into the realm of sequential or video data analysis, particularly relevant in healthcare applications. The main proposition of this research lies in its novel approach to utilizing LSTM networks to analyse time-series data of human physio activities, with a specific focus on rehabilitation exercises targeting low back pain. By leveraging the publicly available KIMORE dataset, containing recordings of various rehabilitation exercises and corresponding participant performance data, the research aims to develop a robust LSTM model capable of automatic evaluation of these activities. The study employs a basic design involving data pre-processing and splitting using Scikit-learn packages, followed by the implementation of an LSTM model using Python and TensorFlow. The key findings highlight the efficacy of LSTM in evaluating participant performance and analysing time-series data, thus contributing to advancements in healthcare analytics. Overall, this research underscores the potential of LSTM networks in healthcare applications, particularly in analysing sequential or time-series data for improved patient care and treatment monitoring.

Keywords: Computer Vision, Machine Learning, Neural Networks

 Conference Details

Session: Poster Session B at Poster Stand 56

Location: Sir Stanley Clarke Auditorium at Wednesday 8th 09:00 – 12:30

Markers: Jay Morgan, Gregory Cheng

Course: BSc Computer Science, 3rd Year

Future Plans: I’m looking for work