The initial hypothesis of the project was: The human visual system is not capable of perceiving everything that happens on a screen. Thus the quality of a graphics rendering can be improved by using an eye tracker to track the gaze point, and concentrate the use of computational resources to render the area around the gaze point and use less resources in the peripheral of the vision, without it being perceptible to the eye. Preliminary tests proved that users were unable to perceive the lower quality rendering when a gaze area, of 20 degrees visual angle, was rendered in high quality at the gaze point tracked by an eye tracker. The gaze area was rendered with global illumination using photon mapping while the outer area was rendered with local illumination using OpenGL. To fully benefit from gaze directed hybrid rendering, a gaze directed photon mapping algorithm was developed. The algorithm guides photons toward the gaze region to achieve a higher photon density in this region. A test of the real-time visual quality revealed that 77 % of the test subjects preferred gaze directed photon mapping instead of traditional photon mapping, and a quantitative analysis of the number of photons in the gaze region showed an increase of up to 85 %. The frame rate of hybrid rendering with gaze directed photon mapping was up to five times faster than rendering the entire screen with global illumination using photon mapping. The conclusion is that gaze directed hybrid rendering with gaze directed photon mapping can be used to achieve computer graphics of better quality at lower computation costs, without loss of visual quality.