In order to provide a technology that can remove noise using an isotropic 3D adaptive filter for 3D images, the target pixel is centered on the medical image including the 3D reconstructed image 40 of the subject. The three-dimensional region of interest 41a-41d is set, and the traveling direction of the structures 42a-42c included in the three-dimensional region of interest is detected based on the distribution of the pixel values included in the three-dimensional region of interest. Set a 3D spatial filter that has a shape along the direction of travel of the body, or a 3D smoothing filter that does not consider the shape of the structure if there is no structure, and use these 3D spatial filters to create a 3D region of interest. Noise removal processing is performed.3次元画像に対し、等方的な3次元適応型フィルタを用いたノイズ除去が行える技術を提供するために、被検体の3次元再構成像40を含む医用画像上において、着目画素を中心とする3次元の関心領域41a～41dを設定し、3次元の関心領域に含まれる画素値の分布を基に、3次元の関心領域に含まれる構造体42a～42cの走行方向を検出し、構造体の走行方向に沿った形状を有する3次元空間フィルタ、構造体がない場合は構造の形状を考慮しない3次元平滑化フィルタを設定し、それら3次元空間フィルタを用いて3次元の関心領域に対しノイズ除去処理を行う。
