In several medical applications as well as human safety evaluation, accurate electromagnetic field exposure assessments are required to identify potential side/adverse effects on humans. Computational human models representing anatomy are commonly used to conduct computational dosimetry studies to assess the in situ electric field for quantitative evaluation due to a limitation in conventional human models. The limitation in conventional human models was due to a limited model resolution (typically a few millimeters), which is attributable to the original resolution of medical images. In particular, the importance of the skin layer is suggested in the research agenda of the international standardization body for human electromagnetic exposure. In this paper, we propose a novel method to improve the accuracy of human head skin modeling, which is applicable even to conventional models. To demonstrate the effect of skin modeling on the computed in situ electric field, computational dosimetry is conducted for uniform magnetic field exposure as well as transcranial magnetic stimulation. Computational results indicate that the in situ electric field for uniform exposure is marginally influenced by the skin thickness and model resolution (up to 5%) for different evaluation metrics used in international safety standards. However, the in situ electric field in the skin during transcranial magnetic stimulation and a simulated electrical shaver (non-uniform field exposure) was affected by 11%, which may be worth discussing for optimal brain stimulation considering the side effects of unintended exposure.