。402· 北京科技大学学报 2006年第4期 Time series similar pattern matching based on wavelet and dynamic time warp- ing OU Wenlong,ZHANG Dezheng,YANG Bingru Infomation Engineering School,University of Science and Technology Beijing,Beijng 100083.China ABSTRACT The paper proposed a dy namic time warping(DTW)indexing and similar matching method of time series based on discrete wavelet transfom,which reduced the dimensionality of time series by dis- crete wavelet transform and constructed multi-dimensional index structure by R-tree.The DTW lower bound and its discrete wavelet transform of query sequence were computed to form a query super-rectangle, thus the similar mat ching in original space based on DTW was converted to that in wavelet transfom space based on Euclidian distance.It w as proved that the method guaranteed no false dismissals and proposed the range query algo rithm and nearest neighbor query algorithm.The result show ed that it was a higher query precision and low er computing cost. KEY WORDS time series;pattern mat ching;dynamic time warping:wavelet transform (上接第35页) Infrared image watershed segmentation based on the preprocess by CNN-PDE bias-anisotropic diffusion filter JU Lei,ZHENG Deling,ZHANG Lei) 1)Information Engineering School.University of Science and Technology Beijing,Beijing 100083.China 2)Department of Communications Engineering.Beijng Electronic Science Technology Institute.Bijng 150027.China 3)Shandng Shengli Vocational Colege.Dongying 257062.China ABSTRACT The watershed algorithm leads to over-segment when it was used to segment an infrared im- age.An adjustable biasanisotropic diffusion filter based on CNN-PDE for smoothing an infrared image was studied.In order to remove the residual noise w hich can not be smoothed by filter,the smoothing coeffi- cient and constrain coefficient were used to threshold the gradient image.The result of watershed segmen- tation of a practical infrared image shows the presented method can restrain over-segmentation effectively. KEY WORDS image segmentation;infrared image processing:cellular neural netw orks;anisot ropic diffu- sion;filterTime series similar pattern matching based on wavelet and dynamic time warp￾ing QU Wenlong, ZHANG Dezheng, Y ANG Bingru Inf ormation Engineering School, Universit y of Science and Technology Beijing, Beijing 100083, China ABSTRACT The paper proposed a dy namic time w arping ( DTW) indexing and similar matching method of time series based on discrete w avelet transfo rm, w hich reduced the dimensionality of time series by dis￾crete w avelet transform and co nstructed multi-dimensional index structure by R ＊-tree.The DTW lower bound and its discrete w avelet transform of query sequence w ere computed to fo rm a query super-rectangle, thus the similar matching in o riginal space based on DTW w as converted to that in w avelet transfo rm space based on Euclidian distance .It w as proved that the method guaranteed no false dismissals and proposed the range query algo rithm and nearest neighbor query algorithm .The result show ed that it w as a higher query precisio n and low er computing cost . KEY WORDS time series ;pattern matching ;dynamic time warping ;w avelet transform ( 上接第 395 页) Infrared image watershed segmentation based on the preprocess by CNN-PDE bias-anisotropic diffusion filter JU Lei 1, 2) , ZHENG Deling 1) , ZHANG Lei 3) 1) Information Engineering School, University of S cience and Technology Beijing, Beijing 100083, China 2) Department of C ommunications Engineering, Beijing Electronic Science Technology Institut e, Beijing 150027, China 3) Shandong S hengli Vocational College, Dongying 257062, China ABSTRACT The w atershed algo rithm leads to over-segment when it w as used to segment an infrared im￾age.An adjustable bias-aniso tropic diffusion filter based on CNN-PDE for smoo thing an infrared image was studied .In order to remove the residual noise w hich can no t be smoothed by filter, the smoo thing coeffi￾cient and constrain coefficient w ere used to threshold the g radient image.The result of w atershed segmen￾tation of a practical infrared image show s the presented method can restrain over-segmentation effectively . KEY WORDS image segmentation ;infrared image processing ;cellular neural netw orks ;anisotropic diffu￾sion ;filter · 402 · 北 京 科 技 大 学 学 报 2006 年第 4 期