about network implement

[HuaminWang]

according to the equation 7, the term that marked by red line uses (1-Tr), however in the implement code that in '1st iteration: Updating X0, S0-->S1' block and 'updating S' block in for loop, the implement is quite different which write as Tr*Tr, is that ok? why implement this code block different as the equation description ? thx

[hongwang01]

according to the equation 7, the term that marked by red line uses (1-Tr), however in the implement code that in '1st iteration: Updating X0, S0-->S1' block and 'updating S' block in for loop, the implement is quite different which write as Tr*Tr, is that ok? why implement this code block different as the equation description ? thx

Thanks for your interest in our work! In our paper, Tr means metal trace; For simplicity, in our code, we adopt Tr as non-metal trace, please refer to https://github.com/hongwang01/InDuDoNet/blob/2587e91e6d9945bb9ca1efef223aa16e7ec56266/deeplesion/Dataset.py#L72

[chenjiachengzzz]

according to the equation 7, the term that marked by red line uses (1-Tr), however in the implement code that in '1st iteration: Updating X0, S0-->S1' block and 'updating S' block in for loop, the implement is quite different which write as Tr*Tr, is that ok? why implement this code block different as the equation description ? thx

Thanks for your interest in our work! In our paper, Tr means metal trace; For simplicity, in our code, we adopt Tr as non-metal trace, please refer to

https://github.com/hongwang01/InDuDoNet/blob/2587e91e6d9945bb9ca1efef223aa16e7ec56266/deeplesion/Dataset.py#L72

您好，那不是应该只有一个 tr 吗，为何实现代码里是 tr乘tr

[hongwang01]

according to the equation 7, the term that marked by red line uses (1-Tr), however in the implement code that in '1st iteration: Updating X0, S0-->S1' block and 'updating S' block in for loop, the implement is quite different which write as Tr*Tr, is that ok? why implement this code block different as the equation description ? thx

Thanks for your interest in our work! In our paper, Tr means metal trace; For simplicity, in our code, we adopt Tr as non-metal trace, please refer to https://github.com/hongwang01/InDuDoNet/blob/2587e91e6d9945bb9ca1efef223aa16e7ec56266/deeplesion/Dataset.py#L72

您好，那不是应该只有一个 tr 吗，为何实现代码里是 tr乘tr

Tr是binary matrix，所以Tr=Tr*Tr，不影响的