基于多样性特征与层次聚类的肺野自动分割算法
Journal: Advances in Computer and Autonomous Intelligence Research DOI: 10.12238/acair.v3i4.17877
Abstract
胸部X射线(Chest X-Ray,CXR)图像中肺野的自动分割对于肺部疾病的筛查与诊断具有重要意义。然而,CXR为投影成像,结构复杂、重叠多、边界差、自动分割难,手动勾画耗时且一致性差,故需鲁棒自动化分割方法。但深度学习黑箱性强、泛化差、计算要求高,传统无监督方法对特征与边界刻画不足。为此,本文提出融合超像素、多样性特征工程与层次聚类的分割新方法。该方法通过超像素分割降低数据维度,结合多尺度滤波、纹理分析与邻域统计增强特征判别能力,经贪心策略筛选特征以提升泛化能力与聚类效率,最终通过层次聚类合并区域并优化边界,提升分割精度与一致性。实验显示,该方法在多个公开数据集上表现优异,IoU达0.8791,优于图像检索加配准(0.8634)及OGD特征结合模糊聚类(0.7921)。
Keywords
超像素分割；多样性特征提取；层次聚类
Full Text
PDF - Viewed/Downloaded: 0 TimesReferences
[1]ARRIETA A B,DIAZ-RODRIGUEZ N,DEL SER J,et al.Explain able artificial intelligence(xai):Concepts, taxonomies, opport
unities and challenges toward responsible ai[J].Information Fusion,2020,58:82-115.
[2]ACHANTA R,SHAJI A,SMITH K,et al.Slic superpixels compa red to state-of-the-art superpixel methods[J]. IEEE Transac tions on Pattern Analysis and Machine Intelligence,2012,34(11):2274-2282.
[3]VAN GINNEKEN B,STEGMANN M B, LOOG M.Segmentation of anatomical structures in chest radiographs using supervised methods:a comparative study on a public database[J].Medical Image Analysis,2006,10(1):19-40.
[4]IBRAGIMOV B, LIKAR B,PERNUš F,et al.Fusing shape infor mation in lung segmentation in chest radiographs[J].Computer ized Medical Imaging and Graphics,2014,38(7):632-641.
[5]XU Y,VAN GINNEKEN B, SHIRAISHI J,et al.An edge-region force guided active shape approach for automatic lung field detection in chest radiographs[J].Medical Image Analysis,2012,16(1):252-264.
[6]LEE J,LEE H,PARK H,et al.A game-theoretic framework for landmark-based image segmentation[J]. IEEE Transactions on Image Processing,2016,25(6):2795-2807.
[7]CANDEMIR S,JAEGER S,PALANIAPPAN K,et al.Lung segment ation in chest radiographs using anatomical atlases with nonr igid registration[J]. IEEE transactions on medical imaging,2014,33(2):577-590.
[8]ZHOU Z,SHIN J,ZHANG Y,et al.Hierarchical lung field seg mentation with joint shape and appearance sparse learning [J].IEEE Transactions on Medical Imaging,2017,36(5):1005-1015.
[9]SHI Y,QI S,ZHU X,et al.Lung segmentation in chest radio graphs by means of gaussian kernel-based fcm with spatial constraints[J].Biomedical Signal Processing and Control,2019,49:167-178.
[10]ROTHER C,KOLMOGOROV V,BLAKE A.Grabcut: Interactive foreground extraction using iterated graph cuts[J].ACM Trans actions on Graphics(TOG),2004,23(3):309-314.
[11]OF RADIOLOGICAL TECHNOLOGY(JSRT)J S,(JRS)J R S.Jsrt (japanese society of radiological technology database)[EB/OL].
[12]WAN AHMAD W S H A M,W ZAKI W M D,AHMAD FAUZI M F.Lung segmentation on standard and mobile chest radiographs using oriented gaussian derivatives filter[J].BioMedical Engineeri ng OnLine,2015,14(20).
unities and challenges toward responsible ai[J].Information Fusion,2020,58:82-115.
[2]ACHANTA R,SHAJI A,SMITH K,et al.Slic superpixels compa red to state-of-the-art superpixel methods[J]. IEEE Transac tions on Pattern Analysis and Machine Intelligence,2012,34(11):2274-2282.
[3]VAN GINNEKEN B,STEGMANN M B, LOOG M.Segmentation of anatomical structures in chest radiographs using supervised methods:a comparative study on a public database[J].Medical Image Analysis,2006,10(1):19-40.
[4]IBRAGIMOV B, LIKAR B,PERNUš F,et al.Fusing shape infor mation in lung segmentation in chest radiographs[J].Computer ized Medical Imaging and Graphics,2014,38(7):632-641.
[5]XU Y,VAN GINNEKEN B, SHIRAISHI J,et al.An edge-region force guided active shape approach for automatic lung field detection in chest radiographs[J].Medical Image Analysis,2012,16(1):252-264.
[6]LEE J,LEE H,PARK H,et al.A game-theoretic framework for landmark-based image segmentation[J]. IEEE Transactions on Image Processing,2016,25(6):2795-2807.
[7]CANDEMIR S,JAEGER S,PALANIAPPAN K,et al.Lung segment ation in chest radiographs using anatomical atlases with nonr igid registration[J]. IEEE transactions on medical imaging,2014,33(2):577-590.
[8]ZHOU Z,SHIN J,ZHANG Y,et al.Hierarchical lung field seg mentation with joint shape and appearance sparse learning [J].IEEE Transactions on Medical Imaging,2017,36(5):1005-1015.
[9]SHI Y,QI S,ZHU X,et al.Lung segmentation in chest radio graphs by means of gaussian kernel-based fcm with spatial constraints[J].Biomedical Signal Processing and Control,2019,49:167-178.
[10]ROTHER C,KOLMOGOROV V,BLAKE A.Grabcut: Interactive foreground extraction using iterated graph cuts[J].ACM Trans actions on Graphics(TOG),2004,23(3):309-314.
[11]OF RADIOLOGICAL TECHNOLOGY(JSRT)J S,(JRS)J R S.Jsrt (japanese society of radiological technology database)[EB/OL].
[12]WAN AHMAD W S H A M,W ZAKI W M D,AHMAD FAUZI M F.Lung segmentation on standard and mobile chest radiographs using oriented gaussian derivatives filter[J].BioMedical Engineeri ng OnLine,2015,14(20).
Copyright © 2025 曾昌剑
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License