The IDEA Experiment envisaged at future $e^+$$e^-$ circular colliders (FCC-ee and CEPC) is currently under design and optimization with dedicated full-simulation investigations. We review the design of the IDEA fully-projective fiber-based dual-readout calorimeter simulation. Particular attention is given to general and fundamental limitations of calorimeters operating at colliders, together with the path for the reconstruction of events with complex topologies. Through the study of the distinction between $\gamma$ and $\pi^0$, we illustrate the outstanding particle identification capabilities given by a millimeter shower sampling coupled to modern convolutional neural networks.