The spatial organization of membrane receptors at the nanoscale has major implications in cellular function and signaling. The advent of super-resolution techniques has greatly contributed to our understanding of the cellular membrane. Yet, despite the increased resolution, unbiased quantification of highly dense features, such as molecular aggregates, remains challenging. Here we describe an algorithm based on Bayesian inference of the marker intensity distribution that improves the determination of molecular positions inside dense nanometer-scale molecular aggregates. We tested the performance of the method on synthetic images representing a broad range of experimental conditions, demonstrating its wide applicability. We further applied this approach to STED images of GPI-anchored and model transmembrane proteins expressed in mammalian cells. The analysis revealed subtle differences in the organization of these receptors, emphasizing the role of cortical actin in the compartmentalization of the cell membrane.