You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi @Zhaozixiang1228, excellent paper and I appreciate for open-souring it. I would like to ask a question about the equation (7), which initially is $\nabla_{f_t}\log p_t(f_t|i,v) = \nabla_{f_t}\log p_t(f_t) + \nabla_{f_t}\log p_t(i,v|f_t)$
For the second part on the right side of the equation $\nabla_{f_t}\log p_t(i,v|f_t)$, which is the likelihood that probably to assess how good the fusion process is. In my mind, it serves somehow like the loss function used in image fusion. Therefore, from the perspective of the loss function, will it make the overall fusion better by fine-tuning the model? I understand that inference-only is awesome, but I would like to know if what I'm saying is even feasible.
The text was updated successfully, but these errors were encountered:
Hi @Zhaozixiang1228, excellent paper and I appreciate for open-souring it. I would like to ask a question about the equation (7), which initially is
$\nabla_{f_t}\log p_t(f_t|i,v) = \nabla_{f_t}\log p_t(f_t) + \nabla_{f_t}\log p_t(i,v|f_t)$ $\nabla_{f_t}\log p_t(i,v|f_t)$ , which is the likelihood that probably to assess how good the fusion process is. In my mind, it serves somehow like the loss function used in image fusion. Therefore, from the perspective of the loss function, will it make the overall fusion better by fine-tuning the model? I understand that inference-only is awesome, but I would like to know if what I'm saying is even feasible.
For the second part on the right side of the equation
The text was updated successfully, but these errors were encountered: