Gait quality assessment using self-organising artificial neural networks.

摘要: In this study, the challenge to maximise potential of gait analysis by employing advanced methods was addressed using self-organising neural networks quantify deviation patients' from normal. Data including three-dimensional joint angles, moments and powers two lower limbs pelvis were used train Kohonen artificial learn an abstract definition normal gait. Subsequently, data patients with problems presented network which quantified quality in form a single curve calculating quantisation error during cycle. A sensitivity involving manipulation variables' weighting able highlight specific causes anatomical location timing wrong patterns. Use can be regarded as extension previously described indices because it measures goodness additionally provides information related underlying deviations.