Text plays an significant role in day-to-day life because of its dissimilarities in text size, font, style, orientation and alignment as well as composite background and rich information, as a consequence automatic text detection in natural scenes has several attractive applications. Though, detecting and recognizing such text is all the time a challenging issue. Several text extraction techniques grounded on edge detection, connected component analysis, morphological operators, wavelet transform, texture features, neural network etc. have been established. This paper contributes comparative analysis of different technique which provides efficient performance.
Improvement of the end-to-end scene text recognition method for “text-to-speech” conversion
