As for the medical application image processing device of this invention, whether or not the local territory is divided the edge which is extracted in the local territory setting section and the edge extraction section which set the local territory which centers the attention in three dimensional model presumption section and two dimensional picture which presume three dimensional model of organism organization on the basis of the edge extraction section and two dimensional picture which extract the edge of two dimensional picture of the image of the organism organization which is input from the medical application image pickup device, pixel by part at least, on the basis of the decision result of the decision section and the decision section which are decided, among the local territoriesIt is the territory where it is not divided by the edge which is extracted, in the edge extraction section, at the same time, making use of three dimensional coordinate data which corresponds to the territory where the attention pixel exists while, it features that it possesses the bulging form sensing station which detects bulging form on the basis of the calculation result of the form feature quantitative calculation section and the form feature quantitative calculation section which calculate the form feature quantity of the attention pixel and.本発明の医療用画像処理装置は、医療用撮像装置から入力される生体組織の像の二次元画像のエッジを抽出するエッジ抽出部と、二次元画像に基づき、生体組織の三次元モデルを推定する三次元モデル推定部と、二次元画像において、注目画素を中心とした局所領域を設定する局所領域設定部と、エッジ抽出部において抽出されたエッジの少なくとも一部により、局所領域が分割されているか否かを判定する判定部と、判定部の判定結果に基づき、局所領域のうち、エッジ抽出部において抽出されたエッジにより分割されない領域であり、かつ、注目画素が存在する領域に対応する三次元座標データを用いつつ、注目画素の形状特徴量を算出する形状特徴量算出部と、形状特徴量算出部の算出結果に基づいて隆起形状を検出する隆起形状検出部と、を有することを特徴とする。
