计算机技术与发展2017，Vol.27Issue(8)：30-36,7.DOI:10.3969/j.issn.1673-629X.2017.08.007
一种图像感兴趣区域提取方法研究
Investigation on Extraction of Interests Region in Image
摘要
Abstract
Concerning extraction of Region Of Interest (ROI) for image,the enhanced Stentiford visual modeling method has the advantages of clear image regions,obvious edges,and high efficiency compared with traditional Stentiford or other visual modeling methods.However,when the image background is relatively complex,regions outside the target would be extracted.Because of requirements on accuracy in actual research,it is necessary to pick out target regions from extracted regions.For this reason,during image processing,an extraction method with multiple new image features has been proposed and the classic FP-Growth algorithm in data mining has been introduced.After the training set have been processed by the enhanced Stentiford visual modeling method,image features such as significant entropy and significant entropy density are extracted and the FP-Growth algorithm is employed to find the association rules between image features and target regions.The obtained rules have been applied to a lot of experimental for verifications.The experimental results show that after the enhanced method,the accuracy of the extracted image regions have been improved significantly which indicates its feasibility and effectiveness.关键词
图像感兴趣区域/Stentiford视觉模型/FP-Growth算法/图像特征Key words
ROI/Stentiford visual model/FP-Growth algorithm/image features分类
信息技术与安全科学引用本文复制引用
王诚,范向阳..一种图像感兴趣区域提取方法研究[J].计算机技术与发展,2017,27(8):30-36,7.基金项目
国家自然科学基金资助项目(61071167) （61071167）
