| 摘要: | 圖像搜尋引擎的使用不如預期，主要是因為平台的搜尋結果、功能與服務的提供無法滿足使用者的需求，此外，介面易用的設計，也影響使用平台的意願。 透過國內外文獻歸納圖像搜尋引擎相關之介面與服務設計要素和使用性原則，整合發展出「圖像搜尋平台服務之使用性設計原則」，包含「應用外界與腦中的知識」、「簡化操作結構」、「系統設計明顯易見」、「正確設計配對關係」、「利用自然或人為的侷限」、「考慮可能的人為錯誤」與「標準化」七個構面，各構面包含相關之介面與服務設計原則要素；再依照原則設計圖像搜尋雛型平台，透過使用者實際操作訪談，驗證平台不管在使用經驗或搜尋結果都能滿足使用者需求；由此得知本研究提出之原則確實能夠做為發展圖像搜尋平台的參考原則。更以「基於內容之圖像檢索」概念為基礎，本研究以「小波」技術實作出以圖像內容搜尋相似圖像的功能，有效降低「以文字為基礎之圖像檢索方式」的問題與限制，更提供使用者不同於以往的圖像搜尋經驗，以更直覺的方式搜尋所需圖像，讓圖像搜尋達到更大的效益。 The usage of image search engine is not as expected, since the searching result, the functions, and the services provided by the image search engine have not fulfilled the user’s needs. Besides, the friendly design of the user interfaces is also one of the factors that influence the user’s willingness of using platforms. By synthesizing from the literatures related to the factors and usability of the interface and service designs within the area of image search (image retrieval), an usability principle for image search platforms including seven dimensions is proposed, which consists of “Use knowledge both in the head and world,” “Simplify the structure of tasks,” “Make things visible,” “Get the mappings right,” “Exploit the power of constraints,” “Design for error,” and “Standardize.” Each dimension includes the factors of related interface and service design principles. Based on the proposed principle, an image search platform was implemented. And through the user interview method, the platform was verified that it fits the user’s needs for enhancing searching result and the experience respectively. Therefore, the proposed usability principle can be truly as a general guideline for developing image search platforms. And based on the Content-Based Image Retrieval (CBIR) concept, a prototype CBIR service was implemented by applying wavelet-based technologies. Thus, the problems and limitations with Text-Based Image Retrieval are reduced effectively. Moreover, a brand-new experience of image search is provided, which lets users search target images more intuitively and makes the best benefit for image search. |
