💡 How to use images.cv datasets
All the things you need to know in order to use images.cv datasets
Folder stracture
After you download and unzip the dataset, you will find a data folder at the main directory,
the data folder contain 3 sub-folder: train, test and val.
Each of those sub-folder will contain more folders acording to your categories that you chose to download.
The name of each folder is combined from the anotations of the images, separated by ' ' (empty space char).
the data folder contain 3 sub-folder: train, test and val.
Each of those sub-folder will contain more folders acording to your categories that you chose to download.
The name of each folder is combined from the anotations of the images, separated by ' ' (empty space char).
Labeling
After you download and unzip the dataset, you will find a data folder at the main directory,
the data folder contain 3 sub-folder: train, test and val.
Each of those sub-folder will contain more folders acording to your categories that you chose to download.
The name of each folder is combined from the anotations of the images, separated by ' ' (empty space char).
the data folder contain 3 sub-folder: train, test and val.
Each of those sub-folder will contain more folders acording to your categories that you chose to download.
The name of each folder is combined from the anotations of the images, separated by ' ' (empty space char).
Images format
The format of all the images is .jpg.
We use a complex compression algorithm for providing you with a well detaild but small size image
We use a complex compression algorithm for providing you with a well detaild but small size image
Image augmentation
If you check the 'Augmentation options' on the download page, will increase your train set randomly by 30% acording to the options that you have selected.
To see which image has been augmented, go to the /train folder inside the zip file and look for these suffix:
To see which image has been augmented, go to the /train folder inside the zip file and look for these suffix:
- Rotate: r
- Noise: n
- Flip: f
- Brightness: b