Text Extraction From Big Data Images

Collection of data sets very large and complex that becomes difficult to be processed using on-hand database management tools or traditional data processing applications is called Big Data. Text information in images of big data serves as important clues for many image-based applications. However, locating text from a complex background with multiple colors is a difficult task. In this paper various streetview images have been taken, that are provided online in the Google API, which acts as the Big data in this paper. The proposed framework in this paper consists of Color based partition method and a mesh concept, which is used to obtain the pixel color values of various locations of an image including the text portion and the non text portion. Now this information collected from various images is used to train the neural network classifier and the classifier is used to detect the text from any new given image. Experiments have been carried on various streetview images and text has been extracted successfully from 75% of the images.