基于LBP 算子的醫(yī)學(xué)圖像檢索方法是目前一個研究熱點，但是目前大多數(shù)算法采用的
是卡方距離計算相似度，針對卡方距離的缺點，本文提出了一種基于LBP 算子與EMD 距離的醫(yī)學(xué)圖像檢索的方法。實驗證明，與卡方距離相比，EMD 距離具有較高的檢準(zhǔn)率和檢全率。
關(guān)鍵詞:醫(yī)學(xué)圖像；LBP 算子；EMD 距離；基于內(nèi)容檢索
Abstract: Nowadays the medical image retrieval approach based on Local Binary
Pattern operators is a hot topic. But most of the approaches are using chi-square
distance, which has inevitable disadvantages. In this paper, a novel medical image
retrieval approach based on Local Binary Pattern and Earth Movers’ Distance is
proposed. The experimental results show that the precision and recall of retrieval
are higher compared with chi-square distance.
Keywords: medical image, LBP operator, EMD, content-based retrieval