Standardize each image to make comparable with the same dimensions.
image_feats
.Predict the category for every test image by finding the training image with most similar features.
Find cluster centers of SIFT features in vocab_size
clusters.
vocab_size
clusters. This gets us the cluster centers.Get the normalized frequencies of words per image.
image_feats
Assign labels to each test feature.
vl_svmtrain
using the train features and labels. This gives us an SVM equation to separate each category from the others (1:all).The lambda in vl_svmtrain
that gave me the best accuracy was 0.00001.
Category name | Accuracy | Sample training images | Sample true positives | False positives with true label | False negatives with wrong predicted label | ||||
---|---|---|---|---|---|---|---|---|---|
Kitchen | 0.540 | Bedroom |
Bedroom |
InsideCity |
Office |
||||
Store | 0.410 | TallBuilding |
Industrial |
TallBuilding |
Forest |
||||
Bedroom | 0.370 | LivingRoom |
LivingRoom |
Industrial |
TallBuilding |
||||
LivingRoom | 0.190 | Bedroom |
InsideCity |
Office |
Mountain |
||||
Office | 0.980 | Kitchen |
LivingRoom |
Kitchen |
Kitchen |
||||
Industrial | 0.480 | Kitchen |
Coast |
InsideCity |
Highway |
||||
Suburb | 0.950 | Industrial |
OpenCountry |
InsideCity |
Store |
||||
InsideCity | 0.660 | Kitchen |
Street |
Industrial |
Street |
||||
TallBuilding | 0.770 | Bedroom |
Store |
Forest |
Mountain |
||||
Street | 0.640 | Industrial |
TallBuilding |
Highway |
TallBuilding |
||||
Highway | 0.810 | Street |
Coast |
Coast |
Coast |
||||
OpenCountry | 0.390 | Bedroom |
Mountain |
Suburb |
Coast |
||||
Coast | 0.830 | OpenCountry |
OpenCountry |
InsideCity |
Highway |
||||
Mountain | 0.840 | Industrial |
Street |
OpenCountry |
Suburb |
||||
Forest | 0.930 | OpenCountry |
TallBuilding |
Mountain |
Mountain |
||||
Category name | Accuracy | Sample training images | Sample true positives | False positives with true label | False negatives with wrong predicted label |