Abstract: A method for realistically combining two separately generated video images, such as of a face and a hairstyle, in such a way that the resulting composite image has a natural appearance. Hairstyle images are captured and processed to include synthesized "fill" hair in addition to real hair. When combined with a face image, the "fill" hair is conditionally displayed, to fill gaps between the hairstyle image and the face image, but not over any portion of a predefined face area. After selection of a hairstyle and combination with the face image, adjustments may be made in position, size and color. A color blending procedure smooths out any sharp edges between hair and face images, and further enhances the natural appearance of the composite image.