The omnipresence of resource-constrained embedded systems makes them critical components. Programmers have to provide strong guarantees about their runtime behavior to make them reliable. Among these, giving an upper bound of live memory at runtime is mandatory to prevent heap overflows from happening. The paper proposes a semi-automatic technique to infer the space complexity of ML-like programs with explicit region management. It aims at combining existing formalisms to obtain the space complexity of imperative and purely functional programs in a consistent framework.