Detecting handwritten digits on a WebCam using TensorFlow and OpenCV

Detecting handwritten digits using mnist and TensorFlow is the "hello world" of Machine Learning. We can make it more interesting by reading handwritten digits on a paper from a webcam. Here's some simple code that used OpenCV and TensorFlow to do just that.

[kaiying]

hi may i know why will need to create another three python files such as Digit recognizer DL LR and NN for this?

[sysadmin]

Can I use a custom dataset, for example a deck of cards that the program will recognize the digits on?

[nadanada-pkbk]

I use OpenCV 4.2.0 and TensorFlow 1.5.0 what I should modify in code in order to be execute correctly?

[md.khaledhasan]

I guess your model has attribute (786, ) which is a 1D matrix and image from cam is (28, 28, 1) which is a 3D matrix.
ValueError: Input 0 of layer sequential is incompatible with the layer: expected axis -1 of input shape to have value 784 but received input with shape [None, 28, 28]

[hameedjasim]

So basically if we copy this code and paste it on tensor flow or open cv. We'll get desired output right?

[krishnavamsiyerrapatruni]

The code isn't working for some reason...displays only the number 6 for any number

[oureibrahimat]

So where is the code, this what we see when we click on the link above { THIS IS NOT THE WEB PAGE YOU ARE LOOKING FOR }