资源简介
三维人脸重建
本文提出了一种基于立体的三维人脸重建方法。该算法基于不同视角下的图像对,对人脸进行深度估计。然后,通过部署深度信息生成面的表面网格。为了提高深度估计的精度,利用了人脸的特征。

代码片段和文件信息
function [F] = face( sigmaMNwin_sizefeatures_mdisparity_feature_pointsdispartity1_refmask_middle)
w=(win_size-1)/2;
gaussians = zeros(MN74);
for i=1:68
x=uint32(features_m(i1));
y=uint32(features_m(i2));
gaussians(y-w:y+wx-w:x+wi) = fspecial(‘gauss‘win_size20);
end
c1 = uint32(features_m(2:) + (features_m(30:)-features_m(2:))/3);
c2 = uint32(features_m(4:) + (features_m(32:)-features_m(4:))/2);
c3 = uint32(features_m(36:) + (features_m(14:)-features_m(36:))/2);
c4 = uint32(features_m(30:) + (features_m(16:)-features_m(30:))/3);
c5 = uint32(features_m(2:) + 2*(features_m(30:)-features_m(2:))/3);
c6 = uint32(features_m(30:) + 2*(features_m(16:)-features_m(30:))/3);
gaussians(c1(2)-w:c1(2)+wc1(1)-w:c1(1)+w69) = fspecial(‘gauss‘win_sizesigma(11));
gaussians(c2(2)-w:c2(2)+wc2(1)-w:c2(1)+w70) = fspecial(‘gauss‘win_sizesigma(11));
gaussians(c3(2)-w:c3(2)+wc3(1)-w:c3(1)+w71) = fspecial(‘gauss‘win_sizesigma(11));
gaussians(c4(2)-w:c4(2)+wc4(1)-w:c4(1)+w72) = fspecial(‘gauss‘win_sizesigma(11));
gaussians(c5(2)-w:c5(2)+wc5(1)-w:c5(1)+w73) = fspecial(‘gauss‘win_sizesigma(11));
gaussians(c6(2)-w:c6(2)+wc6(1)-w:c6(1)+w74) = fspecial(‘gauss‘win_sizesigma(11));
disp=zeros(741);
disp(1:681) = disparity_feature_points;
w1=dispartity1_ref(c1(2)-w:c1(2)+wc1(1)-w:c1(1)+w);
w2=dispartity1_ref(c2(2)-w:c2(2)+wc2(1)-w:c2(1)+w);
w3=dispartity1_ref(c3(2)-w:c3(2)+wc3(1)-w:c3(1)+w);
w4=dispartity1_ref(c4(2)-w:c4(2)+wc4(1)-w:c4(1)+w);
w5=dispartity1_ref(c5(2)-w:c5(2)+wc5(1)-w:c5(1)+w);
w6=dispartity1_ref(c6(2)-w:c6(2)+wc6(1)-w:c6(1)+w);
disp(691) = sum(sum(w1))/sum(sum(w1~=0));
disp(701) = sum(sum(w2))/sum(sum(w2~=0));
disp(711) = sum(sum(w3))/sum(sum(w3~=0));
disp(721) = sum(sum(w4))/sum(sum(w4~=0));
disp(731) = sum(sum(w5))/sum(sum(w5~=0));
disp(741) = sum(sum(w6))/sum(sum(w6~=0));
tot=sum(gaussians3);
for i=1:74
gaussians(::i)=gaussians(::i)./tot;
end
F = zeros(MN);
fo=zeros(741);
for k=1:M
for l=1:N
if mask_middle(kl)==1
fo(:1)=gaussians(kl:);
F(kl) = sum(fo.*disp);
end
end
end
%Smooth
% h=fspecial(‘gauss‘10020);
% F=imfilter(Fh);
%Cost function
%J = sum(sum((dispartity1_ref_out - F).^2));
end
属性 大小 日期 时间 名称
----------- --------- ---------- ----- ----
文件 2278 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\face.m
文件 2298 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\face_mask_extraction.m
文件 2817 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\interF.m
文件 5501219 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\ipcv_project_TURKOGLU.pdf
文件 451 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\k_means.m
文件 9283 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\main.m
文件 1996 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\mesh_create_func.m
文件 391 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master\README.md
目录 0 2019-01-08 16:21 3D_face_reconstruction_from_stereo_pairs-master
----------- --------- ---------- ----- ----
5520733 9
相关资源
- matlab_OFDM调制解调(来自剑桥大学)
- Matlab路面裂缝识别69319
- 高灵敏度GPS接收机MATLAB仿真,附捕获
- 基于MATLAB的质点弹道计算与外弹道优
- 阵列天线的matlab仿真
- MATLAB 经典程序源代码大全
- MATLAB小波软阈值去噪代码33473
- 天线阵的波束形成在MATLAB仿真程序及
- 非线性SVM算法-matlab实现
- 《MATLAB 智能算法超级学习手册》-程序
- 组合导航matlab程序
- 读取txt文件内容matlab代码实现
- Matlab实现基于相关的模板匹配程序
- matlab优化工具箱讲解
- 基于MATLAB的快速傅里叶变换
- 光纤传输中的分布傅立叶算法matlab实
- 基于matlab的图像处理源程序
- matlab 椭圆拟合程序
- 算术编码解码matlab源代码
- optical_flow 光流法 matlab 实现程序
- 引导图像滤波器 Matlab实现
- 分形几何中一些经典图形的Matlab画法
- OFDM系统MATLAB仿真代码
- SVM工具箱(matlab中运行)
- 图像小波变换MatLab源代码
- LU分解的MATLAB实现
- 冈萨雷斯数字图像处理matlab版(第三
- 替代数据法的matlab程序
- 用matlab实现的多站定位系统性能仿真
- 通过不同方法进行粗糙集属性约简m
评论
共有 条评论