PROBLEM TO BE SOLVED: To provide a fundus image processing device capable of correcting a blood vessel attribute on an interactive basis, and appropriately granting a blood vessel attribute with a small work load.SOLUTION: A fundus image processing device includes: blood vessel region emphasizing means 10 for acquiring a blood vessel emphasizing image in which a higher value is set to a fundus image for a pixel with higher blood vessel probability hue conversion means (20, 30, and 40) for converting an RGB value acquired from a pixel corresponding to the fundus image to a hue value for each pixel in the blood vessel region having a value with a predetermined value or more in the blood vessel emphasizing image, and acquiring a hue conversion image blood vessel identification image generation means 60 for granting a temporal arteria attribute and a temporal vein attribute with a prescribed value of the hue value as a boundary in the hue conversion image, and generating a blood vessel identification image in which the attribute granted to each pixel in the blood vessel region can be visually identified and blood vessel identification image correcting means 63 for correcting the blood vessel identification image to the arteria attribute, the vein attribute, the arteria/vein attribute for the pixel included in the blood vessel region in the region based on the region and the attribute indicated to the blood vessel identification image.SELECTED DRAWING: Figure 2COPYRIGHT: (C)2019,JPO&INPIT【課題】 対話形式に血管属性の修正を加え、血管属性を少ない作業負荷で的確に付与する眼底画像処理装置を提供する。【解決手段】 眼底画像に対して、血管確率が高い画素ほど高い値が設定された血管強調画像を得る血管領域強調手段10と、血管強調画像の中で所定の値以上の値をもつ血管領域の各画素に対して、眼底画像の対応する画素から取得したRGB値を色相値に変換し、色相変換画像を得る色相変換手段(20,30,40)と、色相変換画像の中で色相値の所定の値を境に仮動脈属性、仮静脈属性を付与し、血管領域の各画素に付与された属性が視覚的に識別可能な血管識別画像を生成する血管識別画像生成手段60と、血管識別画像に対して指示された領域と属性に基づいて、領域内の血管領域に含まれる画素に対して、動脈属性、静脈属性、動静脈属性に修正する血管識別画像修正手段63を有する。【選択図】 図2
