光电工程
光電工程
광전공정
Opto-Electronic Engineering
2015年
9期
8-13
,共6页
鲁棒特征提取%维数约减%最大间距准则%小样本问题
魯棒特徵提取%維數約減%最大間距準則%小樣本問題
로봉특정제취%유수약감%최대간거준칙%소양본문제
robust feature extraction%dimensionality reduction%maximum margin criterion (MMC)%small sample size problem
现有的最大间距准则(Maximum Margin Criterion,MMC)算法对噪声比较敏感,为了克服这一问题,本文提出了一种基于复数核的鲁棒最大间距准则算法(Robust Maximum Margin Criterion,RMMC).首先通过鲁棒的复数核将样本映射到复数再生核Hilbert空间(Complex Reproducing Kernel Hilbert Spaces,CRKHS),然后在CRKHS空间内实施 MMC 算法.另外,本文也提出了一种求解 RMMC 的高效算法.实验表明,本文算法对于噪声图像有较好的鲁棒性,其识别率较高.
現有的最大間距準則(Maximum Margin Criterion,MMC)算法對譟聲比較敏感,為瞭剋服這一問題,本文提齣瞭一種基于複數覈的魯棒最大間距準則算法(Robust Maximum Margin Criterion,RMMC).首先通過魯棒的複數覈將樣本映射到複數再生覈Hilbert空間(Complex Reproducing Kernel Hilbert Spaces,CRKHS),然後在CRKHS空間內實施 MMC 算法.另外,本文也提齣瞭一種求解 RMMC 的高效算法.實驗錶明,本文算法對于譟聲圖像有較好的魯棒性,其識彆率較高.
현유적최대간거준칙(Maximum Margin Criterion,MMC)산법대조성비교민감,위료극복저일문제,본문제출료일충기우복수핵적로봉최대간거준칙산법(Robust Maximum Margin Criterion,RMMC).수선통과로봉적복수핵장양본영사도복수재생핵Hilbert공간(Complex Reproducing Kernel Hilbert Spaces,CRKHS),연후재CRKHS공간내실시 MMC 산법.령외,본문야제출료일충구해 RMMC 적고효산법.실험표명,본문산법대우조성도상유교호적로봉성,기식별솔교고.
The conventional Maximum Margin Criterion (MMC) is sensitive to noise. To address this problem, a Robust Maximum Margin Criterion (RMMC) based on complex kernel is proposed. The samples are first mapped to the Complex Reproducing Kernel Hibert Spaces (CRKHS), and then the MMC is conducted in CRKHS. Besides, an efficient algorithm for implementing MMC is also proposed in this paper. The experimental results shows that RMMC is robust to the noise images and its recognition rates are higher.
