Characterization of brain infarct lesion in rodent models of stroke is crucial to assess the outcome of stroke therapies and to study the disease pathophysiology. However, so far it has been mostly performed by: 1) histological methods, a time-consuming process that lead to significant flaws and tissue distortion; or 2) via MRI imaging, which is faster, yielding 3D information but at high costs. High-resolution micro-CT imaging became, in the last decade, a simple, fast, and cheaper solution, allowing 3D analysis of samples. Here, we describe that high-resolution micro-CT, either using iodine/OsO4 as contrast agents, can be successfully applied for fine quantification and localization of lesion size and edema volume in preclinical stroke models. We successfully correlated this new approach with the standard histological method TTC. In transient MCAO mouse stroke model, we were able to identify/quantify large lesioned areas (segmented in core and penumbra), up to degenerated finer striatal myelinated fibers in a transient ischemic attack (TIA) mouse model, at different timepoints post-ischemia, through manual and automatic segmentation approaches (deep learning). Furthermore, 3D reconstructions of the whole brain allow for brain atlas co-registration of the specific affected brain areas. Hence, the presented methodology, through iodine/OsO4 micro-CT imaging, constitutes a valuable advance in tools for precise and detailed assessment of stroke outcomes in preclinical animal studies.