基于混合神经网络的图像分类矢量量化方法.PDFVIP

  • 0
  • 0
  • 约9.87千字
  • 约 3页
  • 2019-03-22 发布于山东
  • 举报

基于混合神经网络的图像分类矢量量化方法.PDF

33 6 V o l. 33, l . 6 2006 6 A pp lied Science and T echno logy Jun. 2006 : 1009- 671X ( 2006 06- 0021- 03 李万臣, 王 炼 (哈尔滨工程大学 信息与通信工程学院, 黑龙江 哈尔滨 15000 1 : ( SOFM . , . , ( PCA , , . , SOFM . : ; ; ; : T P 183 : A Im age sorting vector quantization m ethod based on m ixed neural netw ork L IW an- ch en, W ANG L ian ( Schoo l o f In fo rm ation and Communication Eng ineer ing, H arb in Eng ineer ing U n iversity, H arb in 150001, China Abs tract: A trad itional algorithm of mi ge compression is proposed based on the neural newt ork o f SOFM. A nd a new and smi p le sort ing vector quant ization m ethod is put forw ard by ana lyz ing the tradit iona l a lgorithm. T he new m ethod uses edge identificat ion, PCA and SOFM to design the codebook, w hich greatly mi proves the objective index and the subjective visua l mi pression. T he smi u lation show s that the new m ethod is better than the traditiona l SOFM mi - age compression algorithm. K eyword s: edge identificat ion; mi age com pression; vector quan tization; neura l newt ork . , 1 图像分类器的设计 . , 1. 1 BPGA SOFM SOFM . , SOFM , . : , ; , . , . PCA SOFM , PCA [ 1]

您可能关注的文档

文档评论(0)

1亿VIP精品文档

相关文档