A recurrent neural network classifier for Doppler ultrasound blood flow signals

Abstract

The aim of this study is to evaluate the diagnostic accuracy of the recurrent neural networks (RNNs) trained with Levenberg–Marquardt algorithm on the Doppler ultrasound blood flow signals. The ophthalmic arterial (OA) and internal carotid arterial (ICA) Doppler signals were decomposed into time–frequency representations using discrete wavelet transform and statistical features were calculated to depict their distribution. The RNNs were implemented for diagnosis of OA and ICA diseases using the statistical features as inputs. We explored the ability of designed and trained Elman RNNs, combined with wavelet preprocessing, to discriminate the Doppler signals recorded from different healthy subjects and subjects suffering from OA and ICA diseases. The classification results demonstrated that the proposed combined wavelet/RNN approach can be useful in analyzing long-term Doppler signals for early recognition of arterial diseases.