基于邊緣盒與低秩背景的圖像顯著區(qū)域檢測算法
發(fā)布時(shí)間:2019-08-16 14:31
【摘要】:針對現(xiàn)有顯著性區(qū)域邊界不明確和檢測效果魯棒性較差等問題,提出了一種新穎的圖像顯著區(qū)域檢測方法,該方法結(jié)合了邊緣盒粗定位和低秩背景模型細(xì)篩選來提高顯著區(qū)域的檢測性能。首先,對基于邊緣盒的圖像顯著區(qū)域檢測方法進(jìn)行改進(jìn),采用OTSU方法自適應(yīng)計(jì)算邊緣模值的最佳分割閾值,以替代固定分割閾值,降低邊界點(diǎn)檢測誤差;其次,在基于邊緣盒檢測到的可疑顯著區(qū)域上,采用魯棒主成分分析方法獲取圖像的低秩分量,構(gòu)建背景模型,并基于背景差分方法剔除背景區(qū)域,減少顯著區(qū)域的虛檢現(xiàn)象。在PASCAL VOC 2007數(shù)據(jù)集上的實(shí)驗(yàn)結(jié)果表明,提出的方法明顯提高了顯著區(qū)域檢測的精確度和召回率,同時(shí)具有較高的檢測效率。
[Abstract]:In order to solve the problems of unclear boundary of salient region and poor robustness of detection effect, a novel image salient region detection method is proposed, which combines edge box rough location and low rank background model fine screening to improve the detection performance of significant region. Firstly, the image salient area detection method based on edge box is improved, and the OTSU method is used to calculate the optimal segmentation threshold of edge modulus adaptively, in order to replace the fixed segmentation threshold and reduce the detection error of boundary points. Secondly, in the suspicious significant area based on edge box detection, robust principal component analysis (PCA) is used to obtain the low rank component of the image, the background model is constructed, and the background region is eliminated based on the background difference method to reduce the false detection of the significant area. The experimental results on PASCAL VOC 2007 data set show that the proposed method obviously improves the accuracy and recall rate of significant area detection, and has high detection efficiency.
【作者單位】: 江蘇師范大學(xué)計(jì)算機(jī)學(xué)院;太原理工大學(xué)電氣與動力工程學(xué)院;中國礦業(yè)大學(xué)計(jì)算機(jī)科學(xué)與技術(shù)學(xué)院;
【基金】:江蘇省教育科學(xué)“十二五”規(guī)劃課題(C-c/2011/02/010) 江蘇省教育科學(xué)“十二五”規(guī)劃2013年度立項(xiàng)課題(D/2013/02/273)的階段性成果 山西省重大專項(xiàng)項(xiàng)目(20131101029)資助
【分類號】:TP391.41
[Abstract]:In order to solve the problems of unclear boundary of salient region and poor robustness of detection effect, a novel image salient region detection method is proposed, which combines edge box rough location and low rank background model fine screening to improve the detection performance of significant region. Firstly, the image salient area detection method based on edge box is improved, and the OTSU method is used to calculate the optimal segmentation threshold of edge modulus adaptively, in order to replace the fixed segmentation threshold and reduce the detection error of boundary points. Secondly, in the suspicious significant area based on edge box detection, robust principal component analysis (PCA) is used to obtain the low rank component of the image, the background model is constructed, and the background region is eliminated based on the background difference method to reduce the false detection of the significant area. The experimental results on PASCAL VOC 2007 data set show that the proposed method obviously improves the accuracy and recall rate of significant area detection, and has high detection efficiency.
【作者單位】: 江蘇師范大學(xué)計(jì)算機(jī)學(xué)院;太原理工大學(xué)電氣與動力工程學(xué)院;中國礦業(yè)大學(xué)計(jì)算機(jī)科學(xué)與技術(shù)學(xué)院;
【基金】:江蘇省教育科學(xué)“十二五”規(guī)劃課題(C-c/2011/02/010) 江蘇省教育科學(xué)“十二五”規(guī)劃2013年度立項(xiàng)課題(D/2013/02/273)的階段性成果 山西省重大專項(xiàng)項(xiàng)目(20131101029)資助
【分類號】:TP391.41
【相似文獻(xiàn)】
相關(guān)期刊論文 前10條
1 梅躍松;楊樹興;莫波;;一種實(shí)時(shí)抗噪的運(yùn)動區(qū)域檢測算法[J];紅外與激光工程;2006年S4期
2 包瑩瑩;沈夢偉;;基于運(yùn)動區(qū)域檢測的黑板設(shè)計(jì)[J];現(xiàn)代計(jì)算機(jī)(專業(yè)版);2009年08期
3 景慧昀;韓琦;牛夏牧;;顯著區(qū)域檢測算法綜述[J];智能計(jì)算機(jī)與應(yīng)用;2014年01期
4 萬夢玉;張夢琪;;基于游戲中區(qū)域檢測碰撞的改進(jìn)措施[J];計(jì)算機(jī)光盤軟件與應(yīng)用;2014年08期
5 王艷娟;陳曉紅;;圖像感興趣區(qū)域檢測技術(shù)[J];科技資訊;2006年32期
6 王艷娟;陳曉紅;黃曉欣;;圖像感興趣區(qū)域檢測技術(shù)[J];計(jì)算機(jī)與數(shù)字工程;2007年05期
7 陳碩;于曉升;吳成東;陳東岳;;用于場景分類的顯著建筑物區(qū)域檢測[J];上海交通大學(xué)學(xué)報(bào);2011年08期
8 肖創(chuàng)柏;司薇;鄧米克;王首道;;非感興趣區(qū)域時(shí)空分析的運(yùn)動顯著區(qū)域檢測[J];北京工業(yè)大學(xué)學(xué)報(bào);2012年01期
9 蔣鵬;秦小麟;劉s,
本文編號:2527495
本文鏈接:http://www.sikaile.net/kejilunwen/ruanjiangongchenglunwen/2527495.html
最近更新
教材專著