【谷速软件编程】matlab源码-基于哈希编码图像检索.docx

??对于原空间数据，对其中心化(center the dada)后在量化阶段阈值为0，整个编码压缩过程的代码为：XX = XX * randn(size(XX,2),bit);Y = zeros(size(XX));Y(XX=0)=1;??????%原数据中心化后，阈值设为0。大于0编码为1，小于0编码为0Y = compactbit(Y);??%将二进制用十进制表示①表1中对于LSH的投影过程(Projection)是数据独立(data-independent)，对于data-independent，[2]中指出”Generally, data-independent methods need longer codes than data-dependent methods to achieve satisfactory performance”，即要获得比数据非独立方法一样满意的表现，数据独立方法需要更长的编码。实验代码主要部分：%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%Function: this is a geometric illustration of LSH recall precision%curves along with the code length changes%Author: Willard (Yuan Yong English Name)%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%?close?all;clear?variables;clc;load?cifar_10yunchao.mat;X=cifar10;X=X(:,1:end-1);?%Get the recall precision[recall{1,1}, precision{1,1}] = test(X, 16,?LSH);[recall{1,2}, precision{1,2}] = test(X, 24,?LSH);[recall{1,3}, precision{1,3}] = test(X, 32,?LSH);[recall{1,4}, precision{1,4}] = test(X, 64,?LSH);[recall{1,5}, precision{1,5}] = test(X, 128,?LSH);?% Draw the plotfigure; hold?on;grid?on;plot(recall{1, 1}, precision{1, 1},g-^,linewidth,2);plot(recall{1, 2}, precision{1, 2},b-s,linewidth,2);plot(recall{1, 3}, precision{1, 3},k-p,linewidth,2);plot(recall{1, 4}, precision{1, 4},m-d,linewidth,2);plot(recall{1, 5}, precision{1, 5},r-o,linewidth,2);xlabel(Recall);ylabel(Precision);legend(LSH-16 bit,LSH-24 bit,LSH-32 bit,LSH-64 bit,LSH-128 bit,Location,NorthEast);图 1实验数据库是从cifar10中提取出来的320维gist特征，59k作为database，1k作为查询数据。正如图1所示，随着编码位数的增加，PR曲线呈增长趋势，mAP增大。?②PCA-ITQ, PCA-RR, LSH, SKLSH对比实验实验代码主要部分：%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Function: this is a geometric illustration of Draw the recall precision curve%Author: Willard (Yuan Yong English Name)%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%?close?all;clear?variables;clc;load?cifar_10yunchao.mat;X=cifar10;X=X(:,1:end-1);bit=32;?%Get the recall precision[recall{1,1}, precision{1,1}] = test(X, bit,?ITQ);[