An image processing apparatus acquires voxel data obtained by imaging an interior of a living body by a modality and generates a three-dimensional medical image by volume rendering using a ray casting method. The image processing apparatus generates the three-dimensional medical image wherein shade of a surface of an inner wall of an intestinal tract is clearly displayed and an abnormal region invasively developed in the interior of the inner wall is distinguishably displayed based on color information corresponding to voxel data placed at a position shifted from the surface by a predetermined distance.