Dataset description
This dataset contains natural images of size 150x150 for 6 categories (buildings, forest, glacier, mountain, sea, and street). The training set contains 6 folders, around 2300 for each category. The testing set contains 3050 images. The goal is to build a model that can recognize the cateogry of each image in the testing set.
Files description
Scenes training set.rar
- contains the training set
Scenes testing test.rar
- contains the testing set
submit.csv
- a sample submission file in the correct format
names.csv
this file contains the IDs of the six categories
Data fields
Image
- the name of image
Label
- the predicted label