Awesome

image_classification_voting_system

This dataset was provided in a competition organized by Fuse Machines Nepal.

Description: ​The data is generated synthetically which consists of valid as well as invalid voted ballot papers. Data Details​: ​There are 48 classes where 0 class corresponds to the invalid voted ballot papers whereas 1-47 class corresponds to valid ballot papers voted to each political party. Task: ​Create a classifier that classifies a given voted ballot paper to the corresponding class.

The task was completed using convolution neural network with tensorflow