I read a article about ont-shot learning from wiki.
https://en.wikipedia.org/wiki/One-shot_learning
From there, in bayesian framework section, they said that $p(I|I_t,O_{fg})$ = $\int p(I|\theta,O_{fg})p(\theta|I_t,O_{fg}) d\theta$, where $\theta$ is the parameters for $O_{fg}$ (foreground class).
how is this transformation possible?