| 29 | 0 | 182 |
| Downloads | Citas | Reads |
A hierarchical retrieval scheme of the accessory image database is proposed based on textile industrial accessory contour feature and region feature. At first smallest enclosed rectangle~[1] feature (degree of accessory coordination) is used to filter the image database to decouple the image search scope. After the accessory contour information and region information are extracted, the fusion multi-feature of the centroid distance Fourier descriptor and distance distribution histogram is adopted to finish image retrieval accurately. All the features above are invariable under translation, scaling and rotation. Results from the test on the image database including 1,000 accessory images demonstrate that the method is effective and practical with high accuracy and fast speed.
[1]YuYong,ShiPengfei,JournalofShanghaiJiaotongUniversity,19989;32(9):7074
[2]ZhangYujing,Content basedVisionInformationRetrieval,SciencePress,20035,910
[3]WangHuifeng,SunZhengxing,WangJian,JournalofComputerResearchandDevelopment,20025,39(5):513523
[4]MingKueiHu,IEEETransactionsonInformationTheory,19622,8(2):179187
[5]WangBotao,SunJing ao,CaiAnni,JournalofImageandGraphics,20013,6(3):296300
[6]DengshengZhang,GuojunLu,InProc.OfIEEEConferenceonMultimediaandExpo(ICME 01),Tokyo,20018,317320
[7]StefanB ,DanielAK ,HansPK ,etal.IEEETransonKnowledgeandDataEngineering,20002,12(1):4557
[8]JiaYunde,MachineVision,SciencePress,20003,2830
[9]GuoLi,SunXinghua,HuangYuanyuan,JournalofImageandGraphics,200210,7(10):10271031
[10]CaoKui,FengYucai,CaoZhongsheng,JournalofComputer aidedDesign&ComputerGraphics,200110,13(10):906912
Basic Information:
DOI：10.19884/j.1672-5220.2004.03.024
China Classification Code:TS101
Citation Information:
[1]SHEN Wen-zhong YANG JieInstitute of Image Processing and Pattern Recognition, Shanghai Jiaotong University, 200030.Fast Image Retrieval of Textile Industrial Accessory Based on Multi-Feature Fusion[J].Journal of Donghua University (English Edition),2004(03):117-122.DOI:10.19884/j.1672-5220.2004.03.024.