search

Nagakiran1 / 4-simple-steps-in-Builiding-OCR

Optical character recognition (OCR) is process of classification of opti- cal patterns contained in a digital image. The character recognition is achieved through segmentation, feature extraction and classification. Keras Deep learning Network is used at here in recognising the Text characters and OpenCV is used in segmenting the text and Noise normalization.
40 stars 12 forks source link

The problem when characters is very close ? #2

Open NguyenThaiHoc1 opened 5 years ago

NguyenThaiHoc1 commented 5 years ago

Hi Nagakiran

I realized that characters are very close so we should OpenCV or tesseract to object or local segmentation this character usually overlap. You have a solution to this problem. because if you don't, we extracted characters one by one which we use training model in machine learning

Thank for reading Thai Hoc

Nagakiran1 commented 5 years ago

Hi Thai Hoc, Actually OpenCV contours are well in segmenting the characters, if not tesseract trained well in doing that so. If both would not work in your case. I suggest you to build RCNN in classifying letter by letter from a word, because this may reflect in downgrading testing accuracy. Actually i am building that Network to work well in cursive writing. I can provide the solution, once i am done with it.

Regards, Naga.