I didn't consider the cases where we can stack 2 rings on 2 fingers, for example, etc ..
To make it easier on myself, I can first ignore how many rings are making a specific stack, and just consider objects to be arranged on the five fingers. For example, I'll just look at a configuration of ##2, 2, 1## as just ##3## objects to be arranged.
The number of objects I can possible have is ##5, 4, 3## and ##2##, so ##x=5P5a+5P4b+5P3c+5P2d##, where those variables need to be determined.
When we have ##5## objects, this means that I have a ring per finger, so ##a=1##.
When we have ##4## objects, this means that there is a stack of ##2## and ##3## single rings. The number of ways I can gather a stack of ##2## is ##5P2##, so ##b=5P2##.
When we have ##3## objects, then we either have a stack of ##3## rings, and ##2## single rings, or ##2## stacks of ##2## rings, and single ring.
For the case where we have a stack of ##3##, then this is similar to the above case, i.e we have ##5P3## ways of arranging that stack.
Basically, I can represent one configuration of the set of objects to be permuted by ##\{(r_ar_b),(r_cr_d),r_e\}##. At first, I have ##5P2## ways of choosing ##(r_ar_b)##, then I'd be left with ##3## rings, so I can choose ##(r_cr_d)## in ##3P2## ways. Moreover, I'll be doing twice the same permutation eventually since I'd have to permute ##\{(r_ar_b),(r_cr_d),r_e\}## and ##\{(r_cr_d),(r_ar_b),r_e\}##.
Finally, I get ##c=5P3+\frac{5P2\times3P2}{2}##.
When we have ##2## objects, then we either have ##4,1##, or ##3, 2##.
The first case is again like the others, so ##5P4## permutations.
For ##3, 2##, it is like the last case. We have ##\{(r_ar_br_c), (r_dr_e)\}##, so ##5P3\times2P2##.
Thus ##d=5P4+5P3\times2P2##.