Eyes represent the most distinctive features of the human face, while their position and movements are a significant source of information about the cognitive and affective state of humans. Precise eye center localization constitutes a challenging problem in many human-computer interaction applications. In this work, an automatic, non-intrusive method is introduced for the precise eye center localization,based on a modified version of the Fast Radial Symmetry Transform. Specifically,computing the components of the transform from different modalities of the original image, we succeeded to improve the accuracy of the eye center localizer. Experiments were performed in two publicly available face databases, where there is a wide variety of ages, ethnicities, poses, lighting conditions, shadows, presence of occlusions (hair, glasses) and reflections. The obtained results demonstrate the superior performance and enhanced accuracy of this technique against the state-of-the-art methods.