Number recognition
The numerical recognition model was made into CNN using tensorflow using the mnist dataset.
The resulting tensorflow file was converted into a tensorflowlite file and the model was placed inside the Android studio. If you enter a number on Android like that, the value of the number is determined using the tensorflowlite model in it.
When the user inputs a number, the boundary coordinates are obtained using the number coordinates.
Use the boundary coordinates to turn the number over to the adapter. The adapter determines the value by using the machine-learning file of the image of the part.
The numerical value is entered in no value and the nearest figure.
