This dataset consists of 47320 images divided among, Training: 41000 Validation: 3640 Test: 2080
The images are of varying sizes and therefore need to be preprocessed to a uniform size before use. The images are of letters from the Swedish sign language SSL (A-Z).
Training contains 1600 images per class. Validation contains 140 images per class. The test contains 80 images per class.
SOURCES
The data is collected from volunteers.
COLLECTION METHODOLOGY
Collected using Python OpenCV library.
GitHub link to project: https://github.com/gooligang/Final_exam
This contains the training, validation, and test dataset. Each dataset has 26 classes containing images of the Swedish sign language alphabet (A-Z). The training dataset contains 1600 images per class, validation 140, and test 80.
