201245981 六、發明說明： 【發明所屬之技術領域】 [0001] 本發明是有關於一種錯誤偵測方法，特別是有關於印表 機之利用語言模型自動偵測錯誤之方法。 【先前技術】 [0002] 目前在列印環保設計上多半使用縮小及雙面列印等方法 來節省紙張，並利用再生紙與自然材料以減少對於大自 然環境的污染。就算有考量到列印内容的部份，也只有 將列印内容區分為空白區域或資料區域，或是進行初步 〇 的圖文分離後，針對文字或圖片的部份分別作縮小，但 完全沒有使用知識庫或自然語言的方法對列印内容進行 進一步的分析。準此，就省紙及節碳的部份而言，浪費 莫過於印錯再重印。因此，本發明擬以内容檢查之方式 進行錯誤偵測，以避免列印到含有錯誤之資料以節省資 源。 【發明内容】 [0003] Ο 綜合所述，關於習知技術所衍生之問題，本發明目的之 一就是在提供一種印表機之利用語言模型自動偵測錯誤 之方法以解決習知技術之問題。 [0004] 具體而言，本發明之印表機之利用語言模型自動偵測錯 誤之方法包含下列各步驟： 步驟S1 :針對一列印文件之一列印内容進行一文字抽取 之動作。 步驟S2 :針對該列印文件之該列印内容進行一斷詞之動 作。 100115891 表單編號A0101 第3頁/共8頁 1002026650-0 201245981 步驟S3 :利用一語言模型針對該列印文件之該列印内容 進行一錯誤偵測。 [0005] 依上開之闡釋，本發明之印表機之利用語言模型自動偵 測錯誤之方法，其可具有一或多個下述優點： (1) 按本發明之印表機之利用語言模型自動偵測錯誤之方 法，倘若進行錯誤偵測後，該列印文件之列印内容無錯 誤，則列印文件直接被列印出來。反之，若偵測該列印 文件之列印内容有錯誤，則產生一警訊或自動修正該錯 誤。 (2) 利用本發明之印表機之利用語言模型自動偵測錯誤之 方法，可避免紙張以及時間之浪費。 (3) 利用本發明之印表機之利用語言模型自動偵測錯誤之 方法，可導入自然語言處理之方法於週邊產品之開發。 【實施方式】 [0006] 首先陳明，謹遵照如後之相關圖式，茲以說明本發明之 印表機之利用語言模型自動偵測錯誤之方法之主要實施 例。另為便於鈞審理解，所述之實施例中所附之相同元 件係以相同之符號標示以說明之，於此合先述明。 [0007] 謹遵於此，懇請參閱第1圖，其係為本發明之印表機之利 用語言模型自動偵測錯誤之方法之步驟流程圖。誠如第1 圖所示，本發明之印表機之利用語言模型自動偵測錯誤 之方法係適用於印表機之列印文件上，其包含下列步驟 步驟S1 :針對一列印文件之一列印内容進行一文字抽取 之動作。實際上，就列印文件之列印内容而言，常存在 100115891 表單編號A0101 第4頁/共8頁 1002026650-0 201245981 Ο 著圖文混雜的情況，就算列印内容完全是純文字，也會 填充著許許多多標註符號。爰此，在進入錯誤偵測階段 前，所述之文字抽取之動作必須將該列印内容之圖文分 離，並抽取出純文字的部份，以供後續之比較分析。 步驟S2 :針對該列印文件之該列印内容進行一斷詞之動 作。具體而言，中文與英文最大不同處在於中文字詞與 字詞之間並無空白間隔。也就是說，在進行列印内容錯 誤偵測前，必須要有斷詞之機制。故，吾人擬以字詞為 分析之基本單元，並將列印内容之文字串根據字詞的基 本單元做斷詞。201245981 VI. Description of the Invention: [Technical Field of the Invention] [0001] The present invention relates to an error detection method, and more particularly to a method for automatically detecting an error using a language model of a printer. [Prior Art] [0002] At present, most of the environmentally-friendly printing designs use methods such as reduction and double-sided printing to save paper and use recycled paper and natural materials to reduce pollution to the natural environment. Even if you consider the part of the printed content, you can only divide the printed content into blank areas or data areas, or after the preliminary image separation, the text or image parts are reduced separately, but there is no such thing. Further analysis of the printed content using a knowledge base or natural language approach. In view of this, in terms of paper saving and carbon saving, the waste is no more than a mistake. Therefore, the present invention intends to perform error detection by means of content inspection to avoid printing to data containing errors to save resources. SUMMARY OF THE INVENTION [0003] In summary, with regard to the problems derived from the prior art, one of the objects of the present invention is to provide a method for automatically detecting errors by using a language model of a printer to solve the problems of the prior art. . Specifically, the method for automatically detecting errors by the language model of the printer of the present invention comprises the following steps: Step S1: performing a text extraction operation for printing the content of one of the printed documents. Step S2: Perform a word break operation on the printed content of the printed document. 100115891 Form No. A0101 Page 3 of 8 1002026650-0 201245981 Step S3: An error detection is performed on the printed content of the printed document using a language model. [0005] According to the above explanation, the printer of the present invention automatically detects errors by using a language model, which may have one or more of the following advantages: (1) The language of use of the printer according to the present invention The method of automatically detecting the error by the model, if there is no error in the printed content of the printed document after the error detection, the printed document is directly printed. Conversely, if an error is detected in the printed content of the printed document, a warning is generated or the error is automatically corrected. (2) The method of automatically detecting errors using the language model of the printer of the present invention can avoid paper and time waste. (3) The method of automatically detecting errors using the language model of the printer of the present invention can be introduced into the development of peripheral products by the method of natural language processing. [Embodiment] [0006] First of all, Chen Ming, in order to follow the related drawings, will explain the main embodiment of the method for automatically detecting errors by the language model of the printer of the present invention. For the sake of easy understanding, the same components as those in the embodiments are denoted by the same reference numerals, and are described in the foregoing. [0007] In view of this, please refer to FIG. 1, which is a flow chart of the steps of the method for automatically detecting errors by using the language model of the printer of the present invention. As shown in Fig. 1, the method for automatically detecting errors by the language model of the printer of the present invention is applied to a printed document of a printer, which comprises the following steps: S1: printing one of the printed documents The content performs a text extraction action. In fact, in terms of the printed contents of the printed documents, there are often 100115891 Form No. A0101 Page 4 / Total 8 Pages 1002026650-0 201245981 In the case of mixed graphics, even if the printed content is completely plain text, Filled with many more notation symbols. Therefore, before entering the error detection phase, the action of extracting the text must separate the graphic text of the printed content, and extract the pure text portion for subsequent comparative analysis. Step S2: Perform a word break operation on the printed content of the printed document. Specifically, the biggest difference between Chinese and English is that there is no gap between Chinese characters and words. In other words, before the error detection of the printed content, there must be a mechanism for breaking words. Therefore, we intend to use the word as the basic unit of analysis, and the text string of the printed content is broken according to the basic unit of the word.
步驟S3 :利用一語言模型針對該列印文件之該列印内容 進行一錯誤偵測。其中，所述之語言模型係一包含自然 語言模式之語言模型。倘若，進行錯誤偵測後，該列印 文件之列印内容無錯誤，則列印文件直接被列印出來。 反之，若偵測該列印文件之列印内容有錯誤，則產生一 警訊或自動修正該錯誤。當產生警訊號或者修正完該錯 誤後，系統會詢問使用者是否繼續列印。除此之外，本 發明藉由微軟公司之WDK驅動程式進行印表機列印功能 之開發，並在偵測列印内容含有打字錯誤時發出警告訊 息以提醒使用者，進而避免紙張以及時間之浪費。是以 ，利用上述錯誤偵測之機制，可提前偵測列印内容含有 錯誤，進而避免列印含有錯誤内容之文件。不僅如此， 更藉以節省紙張碳粉與列印時間，及提高列印之可靠度 ，並得以邁向永續發展。 [0008] 據此，以上所說明者述僅為較佳實施例之揭示，而非用 100115891 表單編號Α0101 第5頁/共8頁 1002026650-0 201245981 以之限定本發明之主要技術特徵。惟，任何未脫離本發 明之精神與範疇，而進行等效之修改或變更，均應包含 於後所附之申請專利範圍中。 【圖式簡單說明】 [0009] 第1圖係為本發明之印表機之利用語言模型自動偵測錯誤 之方法之步驟流程圖。 【主要元件符號說明】 [0010] S卜S3 :印表機之利用語言模型自動偵測錯誤之方法之步 驟 100115891 表單編號A0101 第6頁/共8頁 1002026650-0Step S3: performing an error detection on the printed content of the printed document by using a language model. Wherein, the language model is a language model containing a natural language pattern. If, after error detection, the printed document has no errors in the printed content, the printed document is directly printed. Conversely, if an error is detected in the printed content of the printed document, a warning is generated or the error is automatically corrected. When a warning signal is generated or the error is corrected, the system asks the user whether to continue printing. In addition, the present invention develops the printing function of the printer by the WDK driver of Microsoft Corporation, and sends a warning message to remind the user when detecting that the printing content contains a typing error, thereby avoiding paper and time. waste. Therefore, by using the above-mentioned error detection mechanism, it is possible to detect in advance that the printed content contains an error, thereby avoiding printing a file containing the wrong content. Not only that, but also to save paper toner and printing time, and improve the reliability of printing, and to achieve sustainable development. Accordingly, the above description is only a disclosure of the preferred embodiment, and instead of 100115891 Form No. 1010101, page 5/8 pages, 1002026650-0 201245981, the main technical features of the present invention are defined. However, any equivalent modifications or alterations that do not depart from the spirit and scope of the invention are intended to be included in the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS [0009] Fig. 1 is a flow chart showing the steps of a method for automatically detecting an error using a language model of the printer of the present invention. [Main component symbol description] [0010] SBu S3: Steps of the printer's automatic detection of errors using the language model 100115891 Form No. A0101 Page 6 of 8 1002026650-0