An image processing apparatus includes a storage unit, a slice image generating unit, a region extracting unit and a tubular structure extracting unit. The storage unit stores a volume image of a three-dimensional region of a subject. The slice image generating unit generates a plurality of slice images corresponding to a plurality of slices each substantially perpendicular to a predetermined reference axis from the volume image. The region extracting unit extracts a target region from the plurality of slice images. The tubular structure extracting unit detects an end point from the extracted region, and extracts a tubular structure based on the end point.