You don't "take the absolute value of an interval".
For set A, f(A) means the set of all values of f(x) for x in A: f(A)= \{f(x)| x\in A\}.
For x between -2 and 0, |x|= -x so f(x)= |x|- 1= -x- 1, a linear function. f(-2)= 2- 1= 1 and f(0)= 0- 1= -1 so for x between -2 and 0, f(x) takes on all values between -1 and 1.
For x between 0 and 3, |x|= x so f(x)= |x|- 1= x- 1, a linear function. f(0)= -1 and f(3)= 2 so for x between 0 and 1, f(x) takes on all values between -1 and 2.
That is, for all x between -2 and 3, f(x) takes on all values between -1 and 2. Notice that -2 is not included in (-2, 3] but f(-2)= 1= f(2) so f((-2, 3])= [-1, 2].
f^{-1}(B) is set of all x values such that f(x)\in B.
Notice that, even though the function f(x)= |x|- 1 is NOT one-to-one and so does NOT have an inverse (that is, f^{-1}(x) is not defined) f^{-1}(-2,3]), applied to a set, is defined. If we look at x< 0, we see that f(x)= |x|- 1= -x- 1= 3 when x= -4. f(x) is never equal to -2 but does go between -1 and 3 (and so is in (-2,3]) for x between -4 and 0. For x> 0, f(x)= |x|- 1= x- 1= 3 when x= 4. Again it does not go down to -2 but for every x between 0 and 4, lies in the set (-2, 3]. That is, f^{-1}((-2, 3])= [-4, 4].