Use model: SVM
total 1692 images
Preprocessing: gray scale => Resize to 64*64
Build 5 classifiers for each attributes.
Validation method:
- Holdout validation (HV)
- Cross validation (CV)
| validation method | HV(train) | HV(test) | CV |
|---|---|---|---|
| Score | .830 | .167 | .162 |
| confusion matrix |  |
 |
| validation method | HV(train) | HV(test) | CV |
|---|---|---|---|
| Score | .649 | .191 | .173 |
| confusion matrix |  |
 |
| validation method | HV(train) | HV(test) | CV |
|---|---|---|---|
| Score | .364 | .177 | .189 |
| confusion matrix |  |
 |
| validation method | HV(train) | HV(test) | CV |
|---|---|---|---|
| Score | .188 | .169 | .172 |
| confusion matrix |  |
 |