Algorithm 2 Image reconstruction from a compressive measurement. |
Input: —measurement vector of length k (it i assumed that , where is the image size in pixels, and that the measurement equation is , where is the measured image, and is the binary measurement matrix) |
Input: — binary array such that at pixels p belonging to region j of map i unless . In further notation, () represents a slice of which corresponds to the i-th map. |
Input: —array of l matrices () with dimensions . Here is the generalized matrix inverse g applied to the measurement matrix after taking the differences of its rows with operator
|
Input: —the initial image reconstruction vector of size (set by us to the initial reconstruction result but may also be filled with constant positive values) |
Input: —learning rate (we took ) |
Output: —vector of size n with the reconstructed image |
function IMAGE RECONSTRUCT(, , , , f ) |
for do
|
for do
|
▹ Loop over maps |
|
▹ Expected pixel sums in sectors of map |
|
▹ Current pixel sums in sectors of map j
|
where else 0 |
|
end for
|
end for
|
return
|
▹ Return the reconstructed image |
end function
|