Open AccessA Multi-Object Feature Selection Based Text Detection and Extraction Using Skeletonized Region Optical Character Recognition in-Text Images
Author(s) -
K R. Sanjuna,
K. Dinakaran
Publication year - 2018
Publication title -
international journal of engineering and technology
Language(s) - English
Resource type - Journals
ISSN - 2227-524X
DOI - 10.14419/ijet.v7i3.6.16009
Subject(s) - artificial intelligence , computer science , pattern recognition (psychology) , segmentation , minimum bounding box , object (grammar) , computer vision , smoothing , edge detection , text detection , feature extraction , image segmentation , object detection , image (mathematics) , image processing
Information or content extraction from image is crucial task for obtaining text in natural scene images. The problem arise due to variation in images contains differential object to explore values like, background filling, saturation ,color etc. text projections from different styles varies the essential information which is for wrong understand for detecting characters.so detection of region text need more accuracy to identify the exact object. To consider this problem, to propose a multi-objective feature for text detection and localization based on skeletonized text bound box region of text confidence score. This contributes the intra edge detection, segmentation along skeleton of object reflective. the impact of multi-objective region selection model (MSOR) is to recognize the exact character of style matches using the bounding box region analysis which is to identify the object portion to accomplish the candidate extraction model.To enclose the text region localization of text resolution and hazy image be well identified edge smoothing quick guided filter methods. Further the region are skeletonized to morphing the segmented region of inter segmentation to extract the text.