Number Recognition of Parts Book Schematics using Convolutional Recurrent Neural Network
Erdal Genç, Hee Ran Shin, Jang‐Sik Park, Jong-Kwan Song · 2018
For recognition of digits, letters, symbols; OCR (Optical Character Recognition) is an important method. In order to allow the computer to edit and read the data which might be handwritten or machine written, the data should be converted to the computer language. According to that, OCR lets computers perceive articles or books. Images can also be read and converted to a text file by using OCR. First of the advantages of OCR, is the improve of the device to work more productively although the quantity of workers is reduced. Secondarily, the storage is efficient. This paper is the benchmark of two state-of-the-art OCR algorithms whose environment is simulated by utilizing modified dataset. The results are displayed in Chapter-4.