Classify black and white images of numbers
classification image recognition digits classics back to basics number
This classic dataset consists of 60K training images and 10K test images of numbers.
Each image is 28x28 grayscale. We performed the usual reshaping to a single line of
784 values before using the data in MNIST database.
Our solution also helps us to achieves great accuracy, without any more processing or
knowledge from the user. Just plug and play.
Results from deployment: