數位內容產業在現今資訊時代是一項重要的產業，同時也是大多數生活的一部份，舉凡媒體溝通、教育學習、育樂或者是資訊傳達等，都與數位內容息息相關。有鑒於數位內容中不良的文字與背景配色，會導致該數位內容被識別或認知上的障礙，本研究應用模糊理論與類神經網路的方法來建立一個推論系統，該系統得以依據輸入的背景顏色來推論一個文字顏色，使得這兩個顏色的搭配讓數位內容文字媒體有高度的視認性。本研究所建立的系統利用模糊理論來推論輸入色的色域歸屬，不同色域的顏色有不同的預測模型，而這些模型是由完成訓練的類神經網路所建構，最終再經過解模糊化的程序來獲得特定的背景顏色。本研究進行兩階段的色彩學實驗，最終以多媒體的形式建立了應用系統，經實際操作確實能有效地推論出高度視認性的配色。

Digital content has become an important industry as information age, which closely connects most aspects with media communication, learning & education, leisure entertainment and information transmission of life at all. For ill match colors with letters and background of digital content will cause obstacles to visibility and legibility. Based on inference system building with fuzzy theory and neural networks, this research sets up a method with colors input by backgrounds to infer the colors from letters. Then contribute to these two parts of digital contents, which attained to a harmonious color combination of highly visibility. Utilizing this inference of neural-fuzzy model by watching different inputs of membership to the hue gamut, each hue gamut reacts to a different predicting model for itself. As it had been trained by neural networks, specific colors of background will obtained by de-fuzzy process. After these two steps of color experiments, a practical user interface of multimedia has been built on an application system. Reiterating tests to the user interface ultimately, valid inference to legibility and visibility of match colors as it used to be.