Table 3 Feature Set Performance for 2-model ergodic HMM framework.
From: On Parsing Visual Sequences with the Hidden Markov Model
Set | States | Recall | Precision |
|
|---|---|---|---|---|
A | 4 | 72.2 | 84.0 | 77.6 |
B | 4 | 79.9 | 82.7 | 81.2 |
C | 12 | 90.1 | 86.4 | 88.2 |
D | 6 | 95.8 | 82.8 | 88.8 |
E | 7 | 93.1 | 83.4 | 87.98 |
F | 8 | 81.5 | 81.7 | 81.6 |
G | 4 | 75.1 | 92.5 | 82.9 |
G | 6 | 97.2 | 77.3 | 86.1 |
H | 4 | 76.4 | 90.8 | 83.0 |
H | 6 | 97.5 | 73.1 | 83.5 |
I | 7 | 94.5 | 76.2 | 84.3 |
J | 4 | 83.4 | 89.8 | 86.5 |
