Philip的笔记

编译、链接均成功，但运行时提示"无法定位程序输入点_glutInitWithExit于动态链接库glut32.dll上"。

这是因为程序运行中调用了旧版本的glut32.dll。解决办法是使用\Cg\bin中的glut32.dll和\Cg\lib中的glut32.lib。、

可以将VC的directories指定以上两个目录，或者把VC的bin和lib中的对应文件替换成Cg中的。

： ）

Have fun!

BTW, Cg的版本是2.0.

I am currently working on reconstruction of BRDF at dense grid with the given sparse samples. I've implemented the Pull-Push method stated in section 6.4 of "A Data-Driven Reflectance Model". I employed i, j, k to denote theta_Difference, Phi_Difference and theta_Half respectively.

The given samples are quite sparse and unevenly scattered. With such samples, I just got unfavorable results. Detailedly, my results are excessively smooth, and mostly higher than the closest given samples.

So here come my questions:

1. Whether the Pull-Push method is no good for reconstructing sparse, unevenly scattered data?
2. Is there any alteration to be taken for reconstructing such data? My implementation is strictly in accordance with what you have stated in "A Data-Driven Reflectance Model", and this is convincing as my colleagues have reviewed my code.

Is it equivalent to adopt interpolation methods, such as bicubic interpolation, for an approximate reconstruction?

Inverse Shade Tree

• 7.8-7.14         PULL-PUSH算法
• 7.15-7.21         单个BRDF的4D-2D
• 7.22-7.29        应用ACLS
• 7.30-8.4        单个BRDF绘制