I have a set of digital images that contain computer printed text in different formats. Kindly suggest three to four machine learning models (preferrably with url) that can recognize the text from input images and save the data in some text file.
I think that the following papers and their methods could be helpful for you:
https://www.mygreatlearning.com/blog/building-a-framework-for-a-machine-learning-algorithm-that-can-recognize-handwriting/
Article A New Deep Learning-Based Handwritten Character Recognition ...
Chapter Classification and Learning Methods for Character Recognitio...
You may use parametric or non-parametric models. I would suggest you to read the attached book and some models are here: https://www.face-rec.org/algorithms/AAM/app_models.pdf
https://www.sciencedirect.com/topics/mathematics/nonparametric-model
https://www.cs.cmu.edu/~dsutherl/papers/cvpr-12.pdf
For Implementation, try following libs which uses different models like. "tesseract". You can also try and implement CNN for character recognition training of model with mnist dataset.
https://github.com/tesseract-ocr/tesseract
https://www.learnopencv.com/deep-learning-based-text-recognition-ocr-using-tesseract-and-opencv/
- Badges
- Science topic