Paper Title
Using Amplitude Spectrum for Gait Features Extraction and Classification
Abstract
Automated gait feature analysis is a precise and objective way to measure gait. Amplitude Spectrum (AS) is
created to extract gait length signal of three types of walking speed (slow, normal and fast speeds). This spectrum shows the
ability for extracting seven gait features that are reduced by using Principal Component Analysis (PCA) technique. In this
paper, the Convolutional Encoder (CE) technique is used to classify the extracted gait features according to the kind of walk
speed. In this classification technique, the Hamming Distance is used to calculate the optimum metric path among the
classes. The obtained results show that 97.1% of the considered parameters are appropriate for classifying between three
kinds of walk speeds.
Keywords- Amplitude spectrum, convolutional encoder, Hamming distance, principle component analysis.