Although CT is generally used for affection-of-the-liver diagnosis of liver cirrhosis, liver cell cancer, etc., many cross-sectional images must be scrutinized visually and a burden is great. Then, the technique of extracting the whole liver domain automatically from an abdomen CT picture as the first step of diagnostic support was developed. After detecting a candidate domain by integrated processing of a bandpass filter and a dynamic outline model first, the validity of the proposal technique was checked by giving the GrabCut method by the evaluation experiment using . clinical picture which extracts the whole liver domain.%肝硬変や肝細胞癌などの肝臓疾患診断にはCTが一般的に用いられているが,多数の断面像を目視で精査せねばならず負担が多大である.そこで診断支援の第一段階として腹部CT画像より肝臓領域全体を自動抽出する手法を開発した.まずバンドパスフィルタと動的輪郭モデルの統合処理にて候補領域を検出した後,GrabCut法を施すことにより肝臓領域全体を抽出する.臨床画像を用いた評価実験により提案手法の有効性を確認した.
展开▼