DocumentCode :
3077848
Title :
Optical Character Recognition for scene text detection, mining and recognition
Author :
Nathiya, N. ; Pradeepa, K.
Author_Institution :
Comput. Sci. & Eng., V.S.B. Eng. Coll., Karur, India
fYear :
2013
fDate :
26-28 Dec. 2013
Firstpage :
1
Lastpage :
4
Abstract :
Text detection in scene images is an important prerequisite for many content-based image analysis tasks. In this method is mainly to identify an accurate and robust method for detecting texts in scene images. A fast and effectual cropping algorithm is designed to extract multi oriented text from an image. The input image is first filtered with connected component approach. Connected component clustering is then used to identify candidate text regions based on the maximum difference. The frame of each connected component helps to separate the different text strings from each other. Then normalize candidate word regions and determine whether each region contains text or not. The scale, skew, and color of each candidate can be estimated from CCs, to develop a text/non text classifier for normalized images. In this techniques not only detect text, it also extracts from the image and recognizes the text in terms of storing the recognized words into a separate file by incorporating several key improvements over traditional existing methods to propose a novel CC clustering based scene text detection method, which finally leads to significant performance improvement over the other competitive methods.
Keywords :
data mining; feature extraction; filtering theory; image classification; object detection; optical character recognition; pattern clustering; text analysis; CC clustering based scene text detection method; component clustering; connected component approach; content-based image analysis tasks; cropping algorithm; image filtering; input image; maximum difference; multioriented text extraction; optical character recognition; scene image; scene text mining; scene text recognition; text classifier; Character recognition; Electronic mail; Feature extraction; Filtering; Image color analysis; Image edge detection; Text recognition; Component Clustering (CC); MSER; Optical Character Recognition (OCR); text classifier; text detection;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computational Intelligence and Computing Research (ICCIC), 2013 IEEE International Conference on
Conference_Location :
Enathi
Print_ISBN :
978-1-4799-1594-1
Type :
conf
DOI :
10.1109/ICCIC.2013.6724165
Filename :
6724165
Link To Document :
بازگشت