Is it possible to compute the peak-to-background ratio for the case of a doublespeak, that is when two peaks are not resolved completely?

So far we have (Figure attached), tying to get the p/B ration for the second peak (peak 1)

1.) Counts under peak 0

2.) Counts under peak 1

3.) Total counts under the combined region

As I understand merely subtracting the 3.) -1.) -2.) will give background counts for both peaks, however only background under peak 1 is required.

Thank You