Research Article
Can a Smartphone Diagnose Parkinson Disease? A Deep Neural Network Method and Telediagnosis System Implementation
Table 1
Time frequency features.
| | Parameter type | Features |
| | Frequency | Jitter (local) | | Parameters | Jitter (rap) | | (Number of features: 5) | Jitter (local, absolute) | Jitter (ppq5)
Jitter (ddp) |
| | Harmonicity | Autocorrelation | | Parameters | Noise-to-harmonic | | (Number of features: 3) | Harmonic-to-noise |
| | Pulse | Number of pulses | | Parameters | Mean period | | (Number of features: 4) | Number of periods | | Standard dev. of period |
| | Amplitude | Shimmer (local) | | Parameters | Shimmer (apq3) | | (Number of features: 6) | Shimmer (local, dB) | | Shimmer (apq5) | | Shimmer (dda) | | Shimmer (apq11) |
| | Pitch | Median pitch | | Parameters | Mean pitch | | (Number of features: 5) | Minimum pitch | | Maximum pitch | | Standard deviations |
| | Voicing | Fraction of locally | | Parameters | unvoiced frames | | (Number of features: 4) | Number of voice breaks | | Degree of voice breaks |
| | Types: 6 | Total features: 26 |
|
|
