Table 1.
Classification results from anterior and posterior classified channels using spectral clustering and peak detection.
| Classification results | S1 | S2 | S3 | S4 | S5 | Average | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ant. | Pst. | Ant. | Pst. | Ant. | Pst. | Ant. | Pst. | Ant. | Pst. | Ant. | Pst. | |
| n (truth; Ant. or Pst.) | 153 | 98 | 124 | 114 | 137 | 116 | 145 | 105 | 100 | 155 | 131.8 | 117.6 |
| n (classified; Ant. or Pst.) spectral clustering | 160 | 91 | 126 | 112 | 131 | 122 | 146 | 104 | 108 | 147 | 134.2 | 115.2 |
| Accuracy spectral clustering | 96.4% | 96.6% | 95.2% | 97.2% | 90.6% | 95.2% | ||||||
| n (classified; Ant. or Pst.) peak detection | 149 | 102 | 106 | 132 | 144 | 109 | 134 | 116 | 127 | 128 | 132 | 117.4 |
| Accuracy peak detection | 96.6% | 92.6% | 86.6% | 94.3% | 89.5% | 91.9% | ||||||
The table compares the channels as anterior or posterior to the central sulcus using the spectral clustering and peak detection method for five subjects. The number of ground truth channels for each subject is shown in row 1. The number of channels classified as anterior or posterior for each subject using spectral clustering and peak detection are given in rows 2 and 4, respectively. Rows 3 and 5 show the final classification accuracy with the average accuracy across all subjects in the last column.