A Simple Approach for Text Segmentation in Images based on Curvelet Transform

Abstract-Text segmentation in image has become an active
research area since recent decades. In this paper we present an
efficient simple method to extract text regions from images
possibly with complex background. The proposed method is to
apply curvelet transform on an image to decompose it into set of
directional sub-bands with texture details captured in different
orientations and scales. Curvelets possess the main features of
wavelets, namely, multi scale and time-frequency localization,
but also offer a high degree of directionality and anisotropy.
Instead of recognizing the edges in only horizontal, vertical &
diagonal directions, the proposed system is designed to
recognize the edges in multi directions taking care of text
orientation. The proposed method is applied on different
samples of images and encouraging results are obtained.