Detecting Text in the Scene Text Image Using Fast Fourier Transform

Abstract

Abstract Text in images express a lot of information for image understanding. Detecting and recognizing text facilitates many important applications. The characteristics of a text with different shapes and contour patterns require special computer vision to detect and recognize when compared to other objects in the image. Text detection is the process of detecting text in an image followed by surrounding text with a rectangular bounding box. It can be performed by using image-based or frequency-based techniques. Many applications require text detection in images ranging from automatic detection of traffic signs that help in the transportation system to assist blind people and others in need. The data of this study were Focused Scene Text with the JPG format. The method used to detect the text in images was Fast Fourier Transform (FFT). The final results of the study showed that 95% of the text could be detected in an image by various camera perspectives and was expected to help develop text recognition applications in an image.