The recent advent of high-resolution peripheral quantitative computed tomography (HR-pQCT) provides new opportunities to measure in vivo human bone microarchitecture. Increasingly, cortical porosity (CtPo) is of particular interest due to its relationship with bone quality and turnover. The two approaches that have emerged to measure CtPo from HR-pQCT are threshold-based and density-based methods, and the purpose of this work was to compare the performance of each against a gold-standard synchrotron radiation micro-computed tomography (SRμCT) measurement. Human cadaveric cortical bone specimens (N=23) were measured by SRμCT and HR-pQCT, and high correlations were found for both methods. The density-based approach had an r2=0.939 (95% confidence interval (CI) of +6.17% to +20.99%) and consistently overestimated porosity as measured by SRμCT, while the threshold-based approach had an r2=0.977 and consistently underestimated porosity (95% CI of -2.60% to -10.76%). The density-based approach is prone to beam hardening artifacts and susceptible to natural variations of tissue mineral density (TMD), but is less affected by motion artifacts that may occur in in vivo scans. The threshold-based method has the advantage that it provides structural information that complements the cortical porosity measure, such as number of pores and connectivity, and can accurately detect the larger pores which are the most relevant to bone biomechanical strength. With the first generation HR-pQCT systems the accuracy of detecting pores larger than 140 μm diameter is excellent (r2=0.983; 95% CI of -4.88% to +2.45%). The accuracy of the threshold-based method will improve as new HR-pQCT systems emerge and provide a robust quantitative approach to measure cortical porosity.