The textual representation of landscapes and other natural settings gives us an insight into the location and can reveal some pertinent details. A challenging problem of considerable practical interest is reading characters in unconstrained scene images. This problem can be observed in medical fields. Scene text identification in a general condition is still a very open and difficult research subject, even though various text identification techniques have been established. Usually, simple images contain text that consists of characters of very similar size and font. In deep scene images, where style attributes will vary, this may not be the case. Thus, End to End Scene Text Understanding using computer vision can be applied to detect text from complex images with complex backgrounds and fonts. Using the text we extracted from the images, we can store it in a document and use it in the future.

Image Processing, Text Recognition, Machine Learning, CNN.