This paper describes Image-to-Device Gamma-Compression Gamut Mapping Algorithm based on the concept of image segmentation. Two image segmentation methods, hue-leaf division and image clustering, are combined with the proposed mapping algorithms. The image color distributions on Chroma-Lightness planes are adjusted to fit to the output device gamut in terms of image's gradation preferred or image's chroma preferred.Three typical mapping methods, (1) Device-to-Device Gamma-Compression, (2) Image-to-Device Gamma-Compression and (3) Clipping Compression, are applied to the segmented images.The psychophysical experiment shows the evaluation results depend on not only the image segmentation methods, but also the image contents and the gamut functions. And, if the image's gradation is preferred, Image-to-Device Gamma-Compression GMA is superior to the conventional methods based on the Device-to-Device concept. If the high-chroma image is preferred, Clipping Compression will be a better choice for use.