How to deblur a image matrix blured by two circulant matrix?

Question

We suppose an image matrix $X\in \mathbb{R}^{n_1\times n_2}$ is blurred by two circulant matrices $\Phi_1 \in \mathbb{R}^{n_1\times n_1},\Phi_2 \in \mathbb{R}^{n_2\times n_2}$. We can observe the blurred matrix $Y$ as: $$Y=\Phi_1 X\Phi_2$$ We suppose the band width of the circulant matrices is $2K$, each non-zero element is generated from a gaussian distribution$N(0,\sigma^2)$. And $\sigma$ is known. How can we recover the original image?