作者: Sina Hassanzadeh , Hossein Pourghassem
DOI: 10.1109/CSPA.2011.5759888
关键词: Classifier (UML) 、 Statistical classification 、 Contextual image classification 、 Pattern recognition 、 Computer science 、 Artificial intelligence 、 Computer vision 、 Object detection 、 Centroid 、 Minimum bounding box 、 Feature extraction 、 Pixel
摘要: In this paper, a novel fast logo detection approach in document images is presented. Logos with separated parts usually can affect the process. To overcome problem, some specifications of logos are considered. Our proposed method divided three main sections. first section, horizontal dilation operator used to merge direction. second simple decision classifier applied for classifying and non-logo. final modifying operation detecting separated-part-logo, which has part, based on two used. These include centroid coordinate intersection each logo's part bounding box. The evaluated public image database international logos. Experimental results show its performance problem.