Our proposed EPVS segmentation challenge focuses on detection of Enlarged Perivascular Spaces (EPVS) in brain MRIs. EPVS is a crucial neuroimage indicator of cerebral vascular disease, linked to conditions like small vessel disease and cognitive impairment risks. Existing EPVS quantification methods, based on visual scoring, are imprecise and inconsistent. A recent study using a 3D U-Net model for automatic EPVS segmentation showed limited effectiveness in older populations, highlighting the need for improved methodologies. EPVS identification is complex due to their varied appearance and the difficulty in distinguishing them from other brain features. The 2021 VALDO challenge addressed these issues by providing a dataset for developing automatic segmentation methods. However, data scarcity (6 fully brain MRI volumes) limited the development of comprehensive models. The winning solution, surprisingly, used a random forest classifier, indicating the potential for advanced data-driven approaches. Our challenge aims to advance machine learning models in this field by providing a larger, more diverse dataset of 60 fully annotated volumetric brain scans from multiple sites and scanners. This dataset includes a range of cognitive profiles, essential for developing robust and generalizable algorithms. Our goal is to foster the creation of effective, reproducible methods for analyzing MR images of the EPVS, aiding early diagnosis and monitoring of cerebral vascular diseases. This challenge, featuring data from various sites, will enhance clinical understanding and treatment of neurovascular and neurodegenerative disorders.