The paper mainly describes Photo OCR [1], automated system software for text- extraction from images or simple character recognition feature. The focus is mainly on reliable text-extraction from cellular phone imagery, for text recognition as input from user similar to speech recognition. Generally the commercial application of OCR does not perform well on this task. The Recent progresses on machine learning has shown improved character isolation classification therefore the paper deals on this progress by displaying complete OCR system software by different techniques by identifying text from uncontrolled conditions like blurred images or from non-readable text documents. We also incorporate various font identification schemes for reading and extracting text from each and every fonts by fixing the boundaries before training the dataset. The approach described is capable of identifying text from images in a variety of challenging imaging conditions where commercial available OCR systems fail, notably in the presence of lighter resolution, high image noise, low contrast and other distortions that take place in an image. The software works with latency that is low and has mean time for processing nearly 650 Ms per image. The software for the Photo-OCR is evaluated public benchmark data-sets for text extraction and performs for all results, taken place through survey. The Photo-OCR system can be used with many applications like Android from Google. © Research India Publications.