Enhance the quality of text based on ML / DL algorithms.

After passing a handwritten document through OCR it generates (i.e. predict ) the text as output. That output is not of good quality. which means OCR doesn't predict 100% accurately. There are lots of wrong words predicted some of the words are out of dictionary words. So the task is to enhance the quality of output text based on ML / DL algorithms. Need to develop ML algorithms such as HMM/SVM etc. to enhance the quality of text . To train the ML Model lots of data is required ( for that take some clean data which is 100 % accurate that may taken from news articles or from google anything and also take some noisy data like I have OCR output which is not 100 % accurate purpose of doing this is to trained ML Model properly ) now once the model will be ready then feed that handwritten text which I have . Let's say we have 5 original input documents (which is 100 % accurate) & we will pass these documents to ML Model then get the output. Then we will pass 5 noisy documents to ML Model & get the output . So by doing this we can estimate the ML Model properties, which means we can know how our ML Model is performing . Then depending on that we can improvise our ML Model. So first read this problem statement carefully then if you feel you have the right skill set to do it . Then Kindly ping me. Good luck.