Develop an End-to-End solution for the Handwritten Digit Recognition System ( MNIST Data : https://keras.io/api/datasets/mnist/ ) . The system should have the following specifications
-
It should have an ML model as the core program API - you can find models ( code ) built on TensorFlow or PyTorch easily and use them.
-
The solution should have a frontend to display results - a simple upload image button, where the user uploads an image from the computer and hits predict, and the solution shows the predicted number with the uploaded image.
-
The solution should have a backend ( database connection ) - for simplicity if I upload an image with the same file name as has been uploaded before, the system should not be predicting, it should simply give the output from a stored answer derived from predicting before, while printing on the screen : “Image predicted before“
-
The final output can be a recorded video explaining the system build and displaying results as required above