You're doing it correctly from what I can see, granted this isn't my specialty by any means!
Personally I think it's easier to understand if you assume you are working with N-1 digits in the register with a single bit at the start that controls the sign. For positive numbers you read it just like a four bit binary with "0" at the start. For negative numbers you read the first bit as "1" which means you subtract the next four bits from "10000" and then turn the result into a negative number.
That means that the complement of a number is always a flip of all bits + 1, so the new number would be a flip of all bits to the left of the "1" farthest to the right; which makes sense because subtracting a number from the same number + 1 in binary will always result in flips for all bits and then add 1 back, again flipping the bits from right to left until you reach the bit where the first "1" was originally because there is now a "0" in that bit to hold the "1" carried from the bit to the right.
For your example, 11011 is the complement of 00101 in a 5-bit register so it looks like you understand the concept! Hopefully this was somewhat helpful.
