scale invariant "shapes_residual"

[beginlu]

Thanks for your great work, help me through the paper. But there is still one problem confuse me. In matlab code, i found you transform the intermediate's "shapes_residual" to the "meanshape_resize" coordinate system, while the latter's bounding box is exactly the same as intermediate's shape. Which means the "shapes_residual" would depends on the image resolution, e.g. same image, if i double its size, the "shapes_residual" would be doubled, and the binary feature still the same. Would this affect the global regression? Why not just transform to the unique param's "meanshape"?

[jwyang]

Hi,

before global regression, shape_residual will be divided by the size of bounding box of intermediate shape. Meanwhile, in the prediction stage, the derived shape residual will times the size of bounding box of intermediate shape. Therefore, if the image is scaled, the corresponding residuals will not change as the binary feature. Hope this helps you.

Best,

On Wed, Dec 24, 2014 at 2:30 AM, beginlu [email protected] wrote:

Thanks for your great work, help me through the paper. But there is still one problem confuse me. In matlab code, i found you transform the intermediate's "shapes_residual" to the "meanshape_resize" coordinate system, while the latter's bounding box is exactly the same as intermediate's shape. Which means the "shapes_residual" would depends on the image resolution, e.g. same image, if i double its size, the "shapes_residual" would be doubled, and the binary feature still the same. Would this affect the global regression? Why not just transform to the unique param's "meanshape"?

— Reply to this email directly or view it on GitHub https://github.com/jwyang/face-alignment/issues/9.

Jianwei Yang Beijing University of Posts and Telecommunications Institute of Automation, Chinese Academy of Sciences 14th floor in Intelligence Building 95 Zhongguancun East Road Haidian District, Beijing 100190, China

杨健伟 北京邮电大学 中国科学院自动化研究所 北京海淀区中关村东路95号智能化大厦14层 邮政编码：100190

[beginlu]

Sorry for my carelessness, it's so obvious, and i just missed the 'normalizing' part. I'm trying to implement it in c++ base on your code, it would make me a deeper understanding. Thx again.