A Review on Various Image-Recognition Types and their Efficiency

Now-a-days images and documents play an important role in storing information. Image processing and text extraction are essential to sustain in our digital society. Text extraction and recognition from images and documents is a key step to build efficient indexing methodology for accessing and retrieving of information from any database. Commonly used in online shopping and augmented reality. To put it simple image recognition provides an extensive range of possibilities and services to categorise and validate documents. Image recognition is not optimized for extracting text and information from images with complex backgrounds and fonts. Still the elements or characters cannot be extracted due to character misalignment or writing style. The purpose of this document is to collect information about various techniques and algorithms used for text extraction from images, documents, and their efficiency. This study has reviewed various deep learning methods such as optical character recognition (OCR) and a hybrid technique by combining region-based and texture-based methods for text extraction.