大部分科技期刊封面包含許多資訊，諸如刊名、出版年月、卷冊數和目次等文字區塊，以及一些商標圖案、背景影像，和圖表等複雜的圖像資料，期望電腦能直接從複雜的封面中辨認出刊名文字是相當困難的；經由圖文分離的過程，萃取出封面文字區塊的封面影像會比較易於處理，進而根據封面文字區塊排列結構，選擇適當、有彈性的結構模式來表示封面的結構特性。在本篇論文中，採用隱藏式馬可夫模型（Hidden Markov Models）表示期刊封面的文字區塊結構；針對每一種科技期刊建構對應的隱藏式馬可夫模型。當一未知的期刊封面輸入時，與資料庫中的模型進行比對，找出最佳的比對模型，用以確認期刊的種類；而封面上的出版年、月、卷冊數和目次等文字區塊亦可從模型中對應的結構狀態找出。

Emphasis of thsi study is put on developing an automatic approach to identifying a given unknown technical journal from its cover page. Since journal's cover pages contain a great deal of information, determining the title of an unknown journal using optical character recognition techniques seems difficult. Using the layout structure of text blocks on the journal's cover page is an effective method for distinguishing one journal from the others. In order to achieve efficient layout-structure comparison, a hidden Markov model (HMM) is used to represent the layout structure of text blocks for each kind of journal. Accordingly, title determination of the unknown journal to each HMM in the database. Besides, from the layout structure of the best matched HMM, we can locate the text blocks of the title, issue date, and contents of the given journal (if they are on the journal's cover page), which will be processed in the future for accomplishing an automatic journal-contents recognition system. Experimental results show the feasibility of the proposed approach.