search

MengyuWang826 / SegRefiner

SegRefiner: Towards Model-Agnostic Segmentation Refinement with Discrete Diffusion Process
Apache License 2.0
137 stars 8 forks source link

作者您好，请问这个公式如何推导的呢？突然出现的转置矩阵不太清楚。 #1

Closed WangzcBruce closed 6 months ago

WangzcBruce commented 6 months ago

image image 用贝叶斯推导公式的话，如上图所示，没有出现转置矩阵欸

MengyuWang826 commented 6 months ago

image image 用贝叶斯推导公式的话，如上图所示，没有出现转置矩阵欸

你图中展示的第8步的结果是带有x{t-1}的，而p(x{t-1}|xt)显然不应该依赖于这一项，所以出现了形式的不同。式(9)的详细推导过程可以在Appendix.A中找到。 The result shown in the 8 step of your diagram includes x{t-1}, while p(x_{t-1}|x_t) should obviously not depend on this term, leading to a discrepancy in the form. A detailed derivation of Equ.(9) can be found in Appendix A.