Abstract We typically measure the radii of transiting exoplanets from the transit depth, which are given by the ratio of cross-sectional areas of the planet and star. However, if a star has dark starspots (or bright regions) distributed throughout the transit chord, the transit depth will be biased toward smaller (larger) values, and thus the inferred planet radius will be smaller (larger) if these are unaccounted for. We reparameterize the transit light curve to account for “self-contamination” by photospheric inhomogeneities by splitting the parameter R p /R ⋆ into two parameters: one for the radius ratio, which controls the duration of ingress and egress, and another which measures the possibly contaminated transit depth. We show that this is equivalent to the formulation for contamination by a second star (with positive or negative flux), and that it is sensitive to time-steady inhomogeneity of the stellar photosphere. We use synthetic light curves of spotted stars at high signal-to-noise to show that the radius recovered from measurement of the ingress/egress duration can recover the true radii of planets transiting spotted stars with axisymmetric spot distributions if the limb-darkening parameters are precisely known. We fit time-averaged high signal-to-noise transit light curves from Kepler and Spitzer of 10 planets to measure the planet radii and search for evidence of spot distributions. We find that this sample has a range of measured depths and ingress durations that are self-consistent, providing no strong evidence for contamination by spots. However, there is suggestive evidence for occultation of starspots on Kepler-17, and that relatively bright regions are occulted by the planets of Kepler-412 and HD 80606. Future observations with the James Webb Space Telescope may enable this technique to yield accurate planetary radii in the presence of stellar inhomogeneities.