本論文提出一個基於知識本體之智慧型影像濾波器(Ontology-based Intelligent Image Filter,OIIF),以基因型模糊影像濾波器(Genetic-based Fuzzy Image Filter,GFIF) [20]為基礎,進一步結合知識本體(Ontology),把各種不同影像雜訊知識以知識本體的架構儲存並提供給GFIF做濾波處理。我們使用的知識本體建構工具為Protégé,論文中將介紹以Protégé建立知識本體的流程。在過濾雜訊之前依影像色調分佈的特徵切割成不同的色度平面,分別處理各平面的濾波結果之後再結合成一張影像為輸出。我們實作出一套系統,將影像濾波處理結果與不同的知名濾波器比較。實驗結果證明基於知識本體之智慧型影像濾波器具有相當優越的去除雜訊能力。
In this paper, an Ontology-based Intelligent Image Filter (OIIF) was proposed, which based on the Genetic-based Fuzzy Image Filter (GFIF) [20]. Ontology structure was constructed of image noise knowledge. The filtering process linked to Ontology and received parameters. Protégé was used to construct Ontology and its operation was introduced in this paper. Before the filtering process, an image was partitioned into many planes via the Luminance Partition. The process partitioned an image according to histogram. Output image is composed of many filtered planes. The performance of OIIF was compared with other famous filters. The experimental results improved that OIIF has the excellent ability of noise removal.