傳統的音樂檢索系統主要在提供使用者特定音樂的查詢(target search)。除此之外，使用者也有類型音樂查詢(category search)的需求。在類型音樂查詢中，該類型的所有音都共同具備使用者所定義的概念(semantic concept)。這個由使用者定義的概念在音樂檢索系統上是主觀的且動態產生的。換句話說，同一使用者在不同情境之下對於同一首音樂可能產生不同的解讀概念。為了動態擷取使用者的概念，讓使用者參與在查詢過程的互動機制是必要的。因此， 我們提出將相關回饋(relevance feedback)的機制運用在以內容為主的音樂查詢系統上，讓系統從使用者的相關回饋中學習使用者的概念，並利用這學習出的概念來幫助音樂查詢。 由於使用者可能從整首音樂或音樂片段兩種角度來判斷該音樂是否具備使用者定義的概念。因此，本論文提出用以片段為主的音樂模型(segment-based modeling approach)將音樂表示成音樂片段的集合。進一步再從整首音樂和片段中擷取特徵。 其次，我們針對該問題提出相關演算法來探勘使用者的概念。該演算法先從相關和不相關的音樂資料庫中個別探勘常見樣式，再利用這些樣式建立分類器以區分音樂的相關性。 最後，我們分析各種系統回饋機制對搜尋效果的影響。Most-positive回傳機制會選擇根據目前系統判斷為最相關的物件。Most-informative機制則是回傳系統無法判斷其相關性的音樂物件。Most-informative 機制的目的在增加每回合系統從使用者身上得到的資訊量。Hybrid 則是中和前兩種機制的優點。本文中，我們模擬並比較各種回傳機制的效能。實驗結果顯示相關回饋機制確實能提升查詢的效果。 / Traditional content-based music retrieval system retrieves a specific music object which is similar to the user’s query. There is also a need, category search, for retrieving a specific category of music objects. In category search, music objects of the same category share a common semantic concept which is defined by the user. The concept for category search in music retrieval is subjective and dynamic. Different users at different time may have different interpretations for the same music object. In the music retrieval system along with relevance feedback mechanism, users are expected to be involved in the concept learning process. Relevance feedback enables the system to learn user’s concept dynamically. In this paper, the relevance feedback mechanism for category search of music retrieval based on the semantic concept learning is investigated. We proposed a segment-based music representation to assist the system in discovering user’s concept in terms of low-level music features. Each music object is modeled as a set of significant motivic patterns (SMP) achieved by discovering motivic repeating pattern. Both global and local music features are considered in concept learning. Moreover, to discover user’s semantic concept, a two-phase frequent pattern mining algorithm is proposed to discover common properties from relevant and irrelevant objects respectively and based on which a classifier is derived for distinguishing music objects. Except user’s feedback, three strategies of the system’s feedback to select objects for user’s relevance judgment are investigated. Most-positive strategy returns the most relevant music object to the user while most-informative strategy returns the most uncertain music objects for improving the discrimination power of the next round. Hybrid feedback strategy returns both of them. Comparative experiments are conducted to evaluate effectiveness of the proposed relevance feedback mechanism. Experimental results show that a better precision can be achieved via proposed relevance feedback mechanism.

抽取出音樂當中的近似重複樣式來做音樂檢索可以減少要比對的資料量，但是使用者若使用沒有重複的旋律來查詢便會有找不到歌曲的情況。另一方面，將音樂分段成phrase可以減少樹狀索引結構的空間，亦可減少查詢處理時間，但是使用者的查詢若是跨越phrase的，也將影響查詢結果。 在本論文中，我們比較了以近似重複樣式與phrase兩種不同的音樂特徵用來做音樂檢索的效果以及效率。根據實驗顯示，使用者的查詢是重複旋律的機會大於單一phrase，所以用近似重複樣式作為音樂查詢比對資料效果是比phrase好的。而在1-D List索引結構下，近似重複樣式的效率也優於phrase。除此之外，本論文也提出了一個新的近似重複樣式抽取方法，實驗證明我們的方法是有效的。 / Extract the approximate repeating pattern from music data will decrease the volumes of music data that need to be tested when music retrieve. If the user’s query is not a repeating melody, it can’t retrieve the music that the user wants correctly. In addition, segment the music by phrase will decrease the space that tree-like index structure need, and also decrease the retrieval processing time. If the user’s query is not a single phrase, it will influence the effectiveness of retrieval. In this thesis, we compare the effectiveness and efficiency of music retrieval methods with two different music features (approximate repeating pattern and phrase). According to experiment results, the probability that user’s query is repeating melody is more than the probability that user’s query is a single phrase. Therefore, we are of the opinion that the effectiveness that use approximate repeating pattern to process retrieval is more prominent than the effectiveness that phrase to process retrieval. Furthermore, the efficiency that use approximate repeating pattern to process retrieval is more outstanding than use phrase under 1-D List index structure. Besides, a new approximate repeating pattern extraction method is proposed. Experiment results show that our approximate repeating pattern extraction method can work correctly.

在本論文中我們提出了一種個人化情緒／情境音樂檢索方法。主要的概念為根據使用者的feedback來找出符合該使用者情緒／情境的音樂在features上所具備的特性，藉此達到個人化的效果。為了更明確表示出音樂的特性，我們利用統計features分布情況的方式來做為音樂的表示法。同時，定義了兩層features自動weighting的方法來決定每個feature在不同情緒／情境下的鑑別度。最後，我們將探討音樂特性與音色對不同的情緒／情境會造成什麼樣的影響，並試著分析音樂與情緒／情境間的關係。 / In this paper, an approach for personalized music retrieval based on emotions / situations is proposed. The main concept is to find out the properties of music that caused the user have emotions / situations responses via the user feedback. And using the user feedback will help us to establish a personalized music retrieval system based on emotions / situations. To represent the music properties clearly, we proposed a new method of music representation based on statistics. And we defined a two-phase features re-weighting method to find out the importance of features in different emotions / situations. At last, we will discuss the influence of music properties and timbre on different emotions / situations, and try to analyze the relationship between the emotions / situations.