Symbol | Description | Symbol | Description |
---|---|---|---|
r ∈ R | Dataset of image patches | t ∈ {0, ⋯, T} | Iteration of ActiveLearn |
Str, Ste | Unlabeled training, testing pools | Φ | Training methodology |
| Eligible samples, annotated samples |
| Samples labeled via Φ at t |
| Fuzzy classifier using | k1,t, k2,t | Number of samples in from ω1, ω2 |
M | Number of votes used to generate | ω1, ω2 | Possible classes of r |
τ | Confidence margin | r ↪ ω 1 | Membership of r in class ω1 |
θ | Classifier-dependent threshold for |
| Number of samples in from ω1, ω2 |
p t (r ↪ ω1) | Probability of observing r ↪ ω1 | N t | Samples added to training set at t |
P Δ | Model confidence |
| Probability of observing samples |
| Accuracy of trained classifier at t |
| Total training cost after T iterations |