What can recognise handwritten check marks and notes on a survey questionnaire? [closed]
I have some questionnaires that are manually filled. Its couple of multiple choice questions & one of the MCQ has a comment section.
People have just scratched (big & small ticks, crossing, underlining, circling - nothing uniform to look for) the response.
I now want to digitize the data, but I am aware of the limitations. So, I am attempting to do what I can. I am thinking for check marks, I just define an area around the choice for MCQ and any form of marking will be considered by comparing it to a blank/unresponded questionnaire. In the same way, identify the written portion for comments and clip it. In the end, get the clipped portion in a webform that a human was just transcribe.
On googling, I find opencv might help me - mine is very messy. But could not find any sample project to figure if it works. I am just starting with exploring the options in opencv. I am just seeing openCV for the first time.
Any tips or suggestions or guidance to speed up my process. Will opencv be able to do what need?
Depending on pen or pencil.
its all pen. Blue & black.
When the blue and black colour becomes white.
how does blue & black become white?
basic "classification" problem. look for examples that classify the MNIST dataset. the same approaches (traditional ML, deep learning) will work for your problem. tensorflow might be of interest to you.