An application of Fourier statistical features in scene text detection

• Published in: 2014 International Conference on Contemporary Computing and Informatics (IC3I) pp. 1154 - 1159
• Main Authors: Vinod, H. C., Niranjan, S. K., Manjunath Aradhya, V. N.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2014
• Subjects: Cameras; Feature extraction; Fourier statistical features; Fuzzy C-means clustering; Image color analysis; Image recognition; Mathematical statistical method; Signal processing algorithms; Statistical analysis; Text recognition
• DOI: 10.1109/IC3I.2014.7019660

Text that appears in images contains important and useful data. Text detection and extraction in images have been applied in many applications. In this paper, we propose n Fourier-Statistical Features in RGB space and Mathematical statistical method for detecting and extracting text in camera images. In RGB space Fourier-Statistical Features is used for detecting text in the image of complex background, contrasting fonts, distinct scripts and different font sizes, In RGB space Fourier transform based features with statistical features and then figured out Fourier-Statistical Features from RGB bands are subject to Fuzzy C-means clustering to classify text pixels from the image background. Classified text pixels of text blocks are determined by inspecting the projection profiles, mathematical statistical method and extract the text part from the image. The suggested approach is examined by carrying on experiments on images of low contrast, complex background, multilingual languages, contrasting fonts, and sizes of text in the image.