TWI223162B - Method and computer system for automatic generation of multimedia WWW news contents from broadcast news video - Google Patents
Method and computer system for automatic generation of multimedia WWW news contents from broadcast news video Download PDFInfo
- Publication number
- TWI223162B TWI223162B TW91105842A TW91105842A TWI223162B TW I223162 B TWI223162 B TW I223162B TW 91105842 A TW91105842 A TW 91105842A TW 91105842 A TW91105842 A TW 91105842A TW I223162 B TWI223162 B TW I223162B
- Authority
- TW
- Taiwan
- Prior art keywords
- news
- patent application
- scope
- video
- content
- Prior art date
Links
Abstract
Description
12231||正替换貞 7月,日 五、發明說明(1) 技術範圍 本發明係有關於將電視新聞視訊内容自動轉 媒體新聞網頁内容之方法及電腦系統裝置。 夕 較早技藝之敘述 自有電視以來,電視新聞媒體製作單位,即 累積新聞視訊(τ V n e w s V i d e 〇 )資料,早期儲存媒汗夕始 不同規格的類比磁帶(匣)。為日後取用方便,、言、此夕^ (,)多以年、月、曰的方式來標記、索引。在這、些二磁π (匣)上擷取需要視訊内容時,則僅能先從選定 ^ π 方式找取碟帶(匣),然後在視訊播放系統中, 的 視的方式找到所需的内容。檢索工作不但費時、費2檢 I年I ί f可能記錯曰期而無法找到所要的視訊内容。 ,f 士數位視訊技術曰漸成熟普及,新聞視訊之 ^ 士走向數位化。數位化儲存技術提供了很方引芬 ,索的功能。一般而言,只要給與每則視訊新聞3及 谷)右干相關的關鍵字(KeyWord),則應用一般 (内 f =就能將所需(有)相關新聞故事從儲存媒體中^取 例如’某公眾人物甲與公眾人物乙在財經 聞故事,使用者可以將「人物甲」及「人物^互^新 經」等珣問字串輸入電腦,來檢索儲存於資料庫 $ 聞内容。只要每則新聞内容之索引關鍵字詞設去 貝料庫中所有符合詢問字串的新聞内容,應都可^ &呈12231 || is being replaced July, July V. Description of the invention (1) Technical scope The present invention relates to a method and computer system device for automatically transferring TV news video content to media news web page content. Evening narratives of earlier techniques Since televisions have been in existence, TV news media production units have accumulated news video (τ V n e w s V i d e) data and stored media of analogue tapes (casks) of different specifications in the early days. For the convenience of future use, the words, eve, and eve ^ (,) are often marked and indexed by the year, month, and day. When capturing video content on these two magnetic π (boxes), you can only find the cassette (box) from the selected ^ π method, and then find the required video in the video playback system. content. The retrieval work is not only time-consuming and expensive, but it takes 2 inspections. I year I ί f may remember the date and cannot find the desired video content. Digital video technology is becoming more mature and popular, and news video is becoming more digital. The digital storage technology provides a very powerful function. Generally speaking, as long as the key words (KeyWord) related to the right stem of each video news 3 are provided, the application (inner f = will be able to take the required (have) related news stories from the storage media ^ for example 'A public figure A and public figure B are in a financial story. Users can enter question strings such as "character A" and "character ^ mutual ^ new sutra" into the computer to retrieve the contents stored in the database. The index key words for each news content are set to all news content in the shell database that matches the query string.
1223162 赞猶·貝 日 五、發明說明(2) 現給使用者。 一™— 因此每則視訊新聞所對應之關鍵字詞是否充足及恰 當,即成為檢索成效之關鍵所在。目前設定關鍵字詞的 方式多以人工兼以電腦程式合併完成[1]。其方法大致有 如下兩種方法: 方法一:完全以人工檢視電視新聞,知悉其内容後設定 關鍵字詞[2 ]; 方法二:根據電視新聞内容,以人工或自動方式選取該 新聞故事(内容)之主播文稿之文字檔(如無文字檔則以適 當方式輸入電腦),以人工或電腦程式做斷詞之方式,在 文字檔中擷取適當關鍵字詞,做為視訊新聞内容之關鍵 鍵字組。 上述方法,對當天或近期之新聞内容之「建立索 引」、「設定關鍵字」、與「設定新聞類別」等工作, 雖然必須使用可觀之人力(電腦技術人力),但尚應可適 用[3 ]。然而對非近期之視訊新聞,由於缺乏可資運用的 主播文稿,前述方法即無法施行。 有謂目前語音辨識技術精進,可應用該技術將主播 播音内容加以辨析來取得主播文字内容[2]。然而放諸世 界最先進語音辨識技術,要將主播播音内容以語音辨識 技術轉成文字檔,似尚未成熟。以目前美國電視新聞媒 體(如三大無線網ABC、NBC、CBS),以辨析主播語音來提1223162 Zanju Bayi V. Description of Invention (2) Now to users. One ™ — Therefore, the adequacy and appropriateness of the keywords for each video news is the key to the effectiveness of the search. At present, the way to set keywords is mostly combined by manual and computer programs [1]. The methods are roughly as follows: Method 1: View the TV news completely manually and set keywords after knowing the content [2]; Method 2: Select the news story (contents manually or automatically) based on the TV news content ) The text file of the anchor manuscript (if there is no text file, enter the computer in an appropriate way), use manual or computer programs to perform word segmentation, and extract the appropriate keywords in the text file as the key to the video news content Block. The above methods, such as "building indexes", "setting keywords", and "setting news categories" for news content on the same day or in the near future, although considerable manpower (computer technical manpower) must be used, but it should still be applicable [3 ]. However, for non-recent video news, the aforementioned methods cannot be implemented due to the lack of available anchor manuscripts. It is said that the current speech recognition technology is advanced, and this technology can be used to analyze the anchor broadcast content to obtain the anchor text content [2]. However, putting the most advanced speech recognition technology in the world, it is not yet mature to convert the content of the anchor broadcast into a text file using speech recognition technology. Based on the current US TV news media (such as the three major wireless networks ABC, NBC, and CBS),
第8頁 1223162 -___ 五、發明明T3):— 一 ' ' - 供螢幕文子稿(ci〇sedCaption),供聽障人士使用的技 術及糸統[22],尚需提供主播所參考的文字稿來達到快 速而正確的辨析。在沒文字稿的幫助情形下,則視語者 發音情況’其語音辨識之正確率大約在5 〇 %〜8 〇 %。 又0謂,在若干國家(特別是非拚音文字國家或地 區),是以螢幕上(On —Screen)文字來幫助聽障人士,因 此可以用光學文字辨識(〇CR)技術將文字辨析出來。這種 辨識方法亦有技術困難:螢幕上文字之解析度低,加上 雜訊影響造成影像品質差,因此文字辨識正確率亦較低 (約在4 0 %〜6 5 % )[ 5 ]。因此,使用前述兩種辨識方法所獲 得的文字串,用來做為關鍵字詞組,將很難做^ 容索引及分類之用。 ^ 本發明之概要Page 8 1223162 -___ V. Invention T3):-a ''-for screen draft (ci0sedCaption), technology and system for the hearing impaired [22], the text referenced by the anchor is also required Manuscript to achieve quick and correct discrimination. In the absence of the help of a text draft, the speech speaker's pronunciation ’accuracy rate of speech recognition is about 50% to 80%. It is also said that in some countries (especially non-Pinyin text countries or regions), on-screen text is used to help the hearing impaired. Therefore, optical text recognition (〇CR) technology can be used to identify the text. This recognition method also has technical difficulties: the resolution of the text on the screen is low, and the image quality is poor due to the influence of noise, so the accuracy of text recognition is also low (about 40% ~ 65%) [5]. Therefore, it is difficult to use the text strings obtained by using the foregoing two recognition methods as keyword phrases for content indexing and classification. ^ Summary of the invention
,照本發明,前述缺點及較早技藝中之問題業 ^,毛明人並依本發明實做雛型,並在網 電腦網站,長期測試與評估。(本案在送智慧 5月發明專利前,相關網站及網址並不公開。;D 發明目標 本發明之目標為提供一種方法及電腦系統According to the present invention, the aforesaid disadvantages and problems in earlier techniques ^, Mao Ming people and prototypes according to the present invention, and long-term testing and evaluation on the computer website. (Before sending the May invention patent in this case, the related websites and web sites were not disclosed. D. Object of the invention The object of the invention is to provide a method and a computer system
何v長,之一整段(例如一小時、半小時)電視新聞朝 (V Video)中,全自動方式剪輯出新聞故事單元, 畫面’並篩選出相關且完整之關鍵字組,供建夕进He v Chang, one of the whole paragraphs (for example, one hour, half an hour) of TV news (V Video), automatically cut out the news story unit, the picture 'and filter out relevant and complete keyword groups for Jianxi Enter
第9頁 1223162Page 9 1223162
五、發明說明(4) 内目不内畫}事目電新鍵聞進快 台大了利大。聞構可 及一許音聞η故一之字關新做而 建容除再,的新建上 引另容語新10聞另類文立視訊接 干内。及外低的容用 索之可播上pt新之分個建電視直 若目存.取聞當往内使 之明構主幕ca訊明及各以述聞以 ,節儲擷新相以訊在 頁發架聞螢ed視發引覽析前新者。外聞架的的是些視且 網本此新將0S個本索流分與視用容此新上容大常這其而 聞 ,將可C1各 立上以以電使内 的序内童率將將, 新 構以亦C中 建路加用將供事 放排目常用可以存 其 架可。 將 以網,更,,故 腦構引 W 加際容時對引聞 電架索 用 容網内同比索新 類此字 並 内在聞。叉及體 分如鍵 , 訊可新類交字媒 及例關 引 視也中分字鍵多 引。構 索 聞統其事文關要 索用建 構。新系取故的立所 訊應來字建引將此擷聞中建尋 。視及,文來索可。時新識、搜 用一法析題識構一統同及辨類來 之供方辨標辨建構系,引所分法 類提作以之以及建網頁索中的方 分為實加上h類為聯網、訊步的 容標同容面,分標腦聞字視一速 較久遠的電視台,其長久以來所播 多以錄影帶保存,用日期之先後來 錄影帶不易長久保存外,其新聞節 用,也常因時日久遠,除了 一些非 部份的新聞視訊内容,在日後的利 因此如利用本發明之系統裝置,即 錄影轉為數位式内容儲存,並且可 出文字標題及摘要。不但易於保 以用關鍵字或分類別等方法來有效V. Description of the invention (4) Inner eyes do not draw in the picture Wen Wenke and Xu Yinwen η therefore a word about new work and new content to build, in addition to the new content, the introduction of the new content of the new 10 Wenwen alternative Wenli video trunk. You can broadcast the new pts and separate TV sets that you can watch right now. Learn from the inside and make it clear the main screen. Ca ’s information and descriptions. The news is published on the page. The outside news is that some of the video and netbooks will be divided into 0S local cables and visually used. This will make the C1 stand up and use the electricity to make the internal child rate. The new construction will also be built in the C road, and will be used for scheduling purposes. Will use the Internet, and more, so the brain structure of W when we add the capacity to the news of the electric cable, the capacity of the network to find the new category and the word. Fork and body are divided into keys, and new types of crosswords and examples can be added to the video. Partial keys are also cited. The structure of the news is related to the use of structure. The news of the new system should be based on this news. Seeing that, Wenlai Soco. New knowledge, search and analysis of the problem by one method, construction of unified identification and identification of supplier-side identification and identification construction systems, quoting the classification of the legal category and the method of building web pages into real and h categories In order to connect to the network, Xunbu ’s standard and face to face, the TV stations with a long history of sub-standards are often watched as videotapes. For the longest time, the videotapes are saved, and it is not easy to save the videotapes for a long time. It is often time-saving, except for some non-partial news and video content. In the future, if the system device of the present invention is used, the video is converted to digital content storage, and text titles and abstracts can be produced. Not only is it easy to keep it effective with keywords or subcategories
第10頁 122316^ 五、發明說明(5) 的檢選出所 本發明 收費機制。 網路網頁瀏 供大眾瀏覽 容便漸對使 較深入的新 載,即設有 瀏覽者可取 本發明 需的 之另 大部 覽者 〇但 用者 聞内 權限 得所 又一 '者,在任 聞查詢所 簡言 容」、厂 引」之方 本發明之 (h 1 )視訊 (h2)類比 (h3)多媒 20, 21,30 (h 4 )數位 (h 5 )數位 60 ) (h 6 )網際 何時間 需新聞 之,本 關鍵字 法及聯 系統裝 輸入裝 視訊轉 體數位 ) 資訊通 資訊儲 新聞 一目 份網 ,原 為尊 設有 容如 及收 需多 目標 任何 資訊 發明 内容。 標為建購一網頁瀏覽 際網路網頁内容為了 則上均無權限管制並 重電視新聞内容之智 權限及收費的機制。 時間久遠以及高品質 費的方法。只有已註 媒體資訊。 為建購產生行動新聞 地點使用無線裝置時 者權限劃 服務多數 以免費方 產權,網 本發明對 視訊内容 冊及願付 瀏覽網供 ,即可瀏 分及 網際 式提 頁内 於比 的下 費的 使用 覽新 乃是一自動產生「視訊新聞内 及自動進行「分類」、「建立索 腦系統裝置。 (1 )所示包括: 網電 置硬 置( 數位視訊處理器(方塊1 1 ) 資訊處理器(個人電腦工作站)(方塊 體設備如圖 方塊1 0 ) 視訊處理器 道 存伺服器(個人電腦工作站 網路介面Page 10 122316 ^ V. Selection Office of the Description of the Invention (5) The charging mechanism of the present invention. The webpages for the public to browse will gradually become more in-depth and new, that is, there are viewers who can access the other large sections required by the present invention. However, users who have the rights within the news can get more. Query the brief description of the contents, "factory quote" of the invention (h 1) video (h2) analog (h3) multimedia 20, 21, 30 (h 4) digital (h 5) digital 60) (h 6) Internet Time needs news, this keyword method and system installation input video rotation digital) Infocom News Store News Yimufenwang. It was originally designed to contain any information invention content that contains tolerance and needs. It is marked as building and purchasing a web page. The Internet web content has no authority to control and pay attention to the intellectual rights and charging mechanism of TV news content. Long time and high quality cost. Only noted media information. For the purpose of generating and purchasing mobile news locations, the use of wireless devices is mostly free of charge. The online invention of the video content booklet and willing to pay for the online supply, you can browse points and Internet-based page-by-page charges. The use of Xinxin is an automatic generation of "in the video news and automatic" classification "," to create a cable system device. (1) shown includes: network power hard (digital video processor (box 1 1) information Processor (personal computer workstation) (block device as shown in Figure 10) Video processor storage server (personal computer workstation network interface
方塊Cube
第11頁 1223162 t 7-^ 五、發明說明(6) 此外, 電腦及網路 本發明 配合本系 瀏覽器。 裝置中有 統,使用者需備有一可上網之個人 則或多則串接在 視訊新聞中 現於網頁上 供做索引及 本發明 (s 1 )新聞視 關的控 起的新 鍵内容 能將文 之用。 成以單 故 (s 2 )語音辨 故事的關 ,同時也 分類整理 系統裝置之資訊 訊分割子系統一 事為個體之視訊 識子系統一將主 制及資 聞視訊 萃取出 字内容 訊處理 故事加 ,並以 中關鍵 系統, 以分析 多媒體 字解析 為能對 ,而將 方式表 出來, 處理子系統包括: 能將多個串接數位新 片段; 播播講語句解析出來 文字 (s3) 影像 (s4 ) (s5) (s 6 ) 台, (s7) (s 8 ) (s9) (si 0 (si 1 稿, 螢幕文字辨識子系統一將螢幕上與内容有關 訊之文字内容 分析子系統; 多媒體新聞内容資料伺服器子系統; 使用者介面子系統一供使用者輸入日期、時 或關鍵字詞來選擇或指定所需新聞内容; 數位新聞内容網頁編輯子系統; 使用者登入子系統; 使用者註冊子系統; )使用者付費驗證子系統; )行動網頁瀏覽子系統。 上述第(s 4 )項所述之「網際網 加以辨識成文字串,做為該視 網際網路文字新聞搜尋及内容Page 11 1223162 t 7- ^ V. Description of the invention (6) In addition, the computer and the Internet cooperate with the browser of this department. The device is unified, and the user needs to have a personal Internet connection or multiple serial links in the video news, which are now indexed on the webpage and the new key content controlled by the news watch of the present invention (s 1) can be used For text. It becomes the key to distinguish stories with single reason (s 2) speech. At the same time, it also sorts and sorts the information and information subsystems of the system devices. It is the individual video information subsystem. One extracts the master system and information videos and extracts the word content. The key system is based on the analysis of the multimedia word analysis as a pair, and the methods are listed. The processing subsystem includes: a plurality of new digital segments can be concatenated; the broadcast sentence is parsed out of the text (s3) image (s4) (s5) (s6) stations, (s7) (s8) (s9) (si0 (si1 draft, screen text recognition subsystem-a text content analysis subsystem that screens content-related information on the screen; multimedia news content Data server subsystem; user interface subsystem-for users to enter date, hour or keywords to select or specify the required news content; digital news content web page editing subsystem; user login subsystem; user registration subsystem System;) user payment verification subsystem;) mobile web browsing subsystem. The “Internet” identified in (s 4) above shall be recognized as a text string as the search and content of the Internet text news
聞分割 ,轉成 的文字 間及電 路文字新聞搜尋及内News segmentation, converted to the text and circuit text news search and internal
第12頁 1223162 5 f 7, _ 五、發明說明(7) 容分析子系統」,為一能根據第(s 6 )項「使用者介面」 所選取或指定之日期、時間及電台資訊,在網際網路上 瀏覽各個文字新聞網站搜尋新聞文稿,並將文稿中擷取 與新聞有關之關鍵詞的一個子系統,同時再利用第(s 2 ) 及第(s 3 )項所述之子系統萃取到的新聞文字串,來比對 在網站上搜尋到的相關新聞之文字稿,如此取得之文字 内容除了可以建立本系統之新聞網頁與網際網路文字新 聞網頁間相互關聯的鏈結(L i n k )外,亦可對(s 2 )及 (s3)項所辨識出不完整之片斷文字串有增補及加強之功 用。第(s 7 )項所述之「數位新聞内容網頁編輯子系’ 統」,此軟體系統主要供專業人員在本系統完成視訊轉 換、文字辨識、關鍵字擷取及内容分類後,並可用人工 檢視的方式,進行快速而有效訂正網頁内容中可能的錯 誤。為了尊重智產權,第(s8) (s9)及(slO)項所述 之使用者「登入」註冊及付費驗證子系統旨在服務一般 網際網路瀏覽大眾外,對新聞資訊有深入需求者,提供 較好的服務。 近年來無線網路日益普及,利用行動裝置,來瀏覽 網際網路資訊,將更為便利,第(s 1 1 )項行動網頁瀏覽 子系統的目的是將傳統新聞網頁内容,自動轉成行動網 頁内容,便於行動裝置使用者,在任何時間、任何地點 均可方便瀏覽最想要知曉的新聞故事。 有關本發明進一步的特性說明,請參閱圖式之簡單敘述 及詳細敘述。Page 12 1223162 5 f 7, _ V. Description of the invention (7) Content analysis subsystem "is a date, time, and station information that can be selected or specified according to (s 6)" User Interface ". Browse various text news websites on the Internet to search for news articles, and extract a sub-system of keywords related to news from the manuscript, and then use the subsystems described in (s 2) and (s 3) to extract News text string to compare the texts of related news searched on the website. In addition to the text content obtained in this way, in addition to establishing a link between the system's news page and the Internet text news page (L ink) In addition, it can also supplement and strengthen the incomplete fragment text strings identified in (s 2) and (s3). The "Digital News Content Web Page Editing Subsystem" described in item (s 7). This software system is mainly used by professionals to complete video conversion, text recognition, keyword extraction, and content classification in this system. Viewing method to quickly and effectively correct possible errors in web content. In order to respect intellectual property rights, the user “login” registration and payment verification subsystem described in (s8) (s9) and (slO) are designed to serve the general Internet browsing public, and those who have in-depth demand for news information, Provide better service. In recent years, wireless networks have become increasingly popular. It is more convenient to use mobile devices to browse Internet information. The purpose of the (s 1 1) mobile web browsing subsystem is to automatically convert the content of traditional news web pages into mobile web pages. Content for mobile device users and easy access to the news stories they want to know at any time, anywhere. For further feature descriptions of the present invention, please refer to the brief description and detailed description of the drawings.
第13頁 1223162 五、發明說明(8) 參考文獻 [1 ] Yasuo Ariki,丨丨 Multimedia Technologies for Structuring and Retrieval of TV News 丨丨,New Generation Computing Vo1. 18, No. 4, pp. 341-358, 2 0 0 0.Page 13 1223162 V. Description of the Invention (8) References [1] Yasuo Ariki, 丨 丨 Multimedia Technologies for Structuring and Retrieval of TV News 丨 丨, New Generation Computing Vo1. 18, No. 4, pp. 341-358, 2 0 0 0.
[2]Y. L. Chang, W. Zeng, I . Kamel, and R. Alonso MIntegrated image and speech analysis for content-based video indexing,M in Proc. of Multimedia, pages 3 0 6 〜3 1 3, Sept. 1 9 9 6.[2] YL Chang, W. Zeng, I. Kamel, and R. Alonso MIntegrated image and speech analysis for content-based video indexing, M in Proc. Of Multimedia, pages 3 0 6 ~ 3 1 3, Sept. 1 9 9 6.
[3 ] F a s T V, http://www.fastv.com [4 ] E r i c Fleischman,”Advanced Streaming Format (ASF) Specification,n Internet Draft, February 26, 1998.[3] Fas T V, http://www.fastv.com [4] E r i Fleischman, "Advanced Streaming Format (ASF) Specification, Internet Draft, February 26, 1998.
[5]Hsin-Chia Fu and Yeong Yuh Xu, 1丨 Multi-linguistic Handwritten Character Recognition by Bayesian Decision-based Neural Networks," in IEEE Transactions On Signal Proc., Vol.46, No. 10, pp. 2781-2789.[5] Hsin-Chia Fu and Yeong Yuh Xu, 1 丨 Multi-linguistic Handwritten Character Recognition by Bayesian Decision-based Neural Networks, " in IEEE Transactions On Signal Proc., Vol. 46, No. 10, pp. 2781- 2789.
第14頁 :31%正替 降年1月5日 五、發明說明(9) [6] Hsin-Chia Fu,Y. Y. Xu and Η. Y. Chang, 丨,Recognition of Similar Handwritten Chinese Characters by Self-growing Probabilistic Decision-based Neural Networks," in International Journal of Neural Systems, Vo 1. 9, No. 6 (December, 1999), ρ·ρ. 545-561· [7] H. C. Fu,P. S. Lai, R. S. Lou, H. -T. Pao, ‘‘Face Detection and Eye Localization by Neural Network Based Color Segmentation’’ in Proc. of NNSP’2000, Sydney, Australia, 11-13 December 2 0 0 0.Page 14: 31% is replaced on January 5th. 5. Description of the invention (9) [6] Hsin-Chia Fu, YY Xu and Y. Y. Chang, 丨, Recognition of Similar Handwritten Chinese Characters by Self-growing Probabilistic Decision-based Neural Networks, " in International Journal of Neural Systems, Vo 1. 9, No. 6 (December, 1999), ρ · ρ. 545-561 · [7] HC Fu, PS Lai, RS Lou, H. -T. Pao, `` Face Detection and Eye Localization by Neural Network Based Color Segmentation '' in Proc. Of NNSP'2000, Sydney, Australia, 11-13 December 2 0 0 0.
[8] H. C· Fu, H.Y· Chang, Υ·Υ· Xu, Η·Τ· Pao, ‘‘User Adaptive Handwriting Recognition by Se1f-growing Probabilistic Decision-based Neural Networks, in IEEE Transactions on Neural Networks, Vo1. 11, No. 6, Nov. 2 0 0 0.[8] H. C · Fu, HY · Chang, Υ · Υ · Xu, Η · Τ · Pao, '' User Adaptive Handwriting Recognition by Se1f-growing Probabilistic Decision-based Neural Networks, in IEEE on Neural Networks, Vo1 . 11, No. 6, Nov. 2 0 0 0.
[9 ] B o r i v o j e Furht, ’’Multimedia systems and techniques," Kluwer Academic, 1996. 10]F r e d Halsall, "Multimedia Communications",[9] B o r i v o j e Furht, ’’ Multimedia systems and techniques, " Kluwer Academic, 1996. 10] F r e d Halsall, " Multimedia Communications ",
第15頁 1223162 翁if h) ·ν ^ 、-· 五、#明說明(10)Page 15 1223162 Weng if h) · ν ^,-· Five, # 明 说明 (10)
Addison-Wesley 2001 [1 1 ] Barry G. Haskell, e t a 1. , "Digital video: An introduction to Mpeg-2”, Chapman and Hall, 1999 [12 ] Q i an Huang, Zhu Liu, Aaron Rosenberg, M Automated semantic structure reconstruction and representation generation for broadcast news, n in Storage and Retrieval for Image and Video Databases, Proc. SP I E 3 6 5 6, ( 1 9 9 9 ).Addison-Wesley 2001 [1 1] Barry G. Haskell, eta 1., " Digital video: An introduction to Mpeg-2 ", Chapman and Hall, 1999 [12] Q i an Huang, Zhu Liu, Aaron Rosenberg, M Automated semantic structure reconstruction and representation generation for broadcast news, n in Storage and Retrieval for Image and Video Databases, Proc. SP IE 3 6 5 6, (1 9 9 9).
[13] Qian Huang, Zhu Liu, Aaron Rosenberg, David[13] Qian Huang, Zhu Liu, Aaron Rosenberg, David
Gibbon, Behzad Shahraray, "Automated Generation of News Content Hierarchy By Integrating Audio, Video, and Text Information," Proc. IEEE International Conference On Acoustics, Speech, and Signal Processing, Phoenix, Arizona, March 15-19, 1999.Gibbon, Behzad Shahraray, " Automated Generation of News Content Hierarchy By Integrating Audio, Video, and Text Information, " Proc. IEEE International Conference On Acoustics, Speech, and Signal Processing, Phoenix, Arizona, March 15-19, 1999.
[14] In formed i a, http://www. informedia. cs. emu. edu [15 ] U r i I urge 1 , R a 1 f Meermeier, Stefan Eickeler and Gerhard Rigoll, ’丨 New Approaches to Audio -Visual Segmentation of TV News for Automatic[14] In formed ia, http: // www. Informedia. Cs. Emu. Edu [15] U ri I urge 1, R a 1 f Meermeier, Stefan Eickeler and Gerhard Rigoll, '丨 New Approaches to Audio -Visual Segmentation of TV News for Automatic
第16頁 12231· 五、發明說明(11)Page 16 12231 · V. Description of the invention (11)
Topic Retrieval’丨, in IEEE In t. Conference on Acoustics, Speech, and Signal Processing (ICASSP), Salt Lake City , Utah, May 2001 [16]Microsoft, http://www.niicrosoft.com/ms.htm [1 7 ] Zhu Liu and Qian Huang, "Classification of audio events in broadcast news,” in Proc. of IEEE Workshop in Multimedia Signal Processing,Topic Retrieval '丨, in IEEE In t. Conference on Acoustics, Speech, and Signal Processing (ICASSP), Salt Lake City, Utah, May 2001 [16] Microsoft, http://www.niicrosoft.com/ms.htm [ 1 7] Zhu Liu and Qian Huang, " Classification of audio events in broadcast news, "in Proc. Of IEEE Workshop in Multimedia Signal Processing,
December 1998.December 1998.
[18]Yasuyuki Nakajima,Yang Lu, Masaru Suganno, Akio Yoneyama, H i romasa Yanagihara, and Akira Kurematsu, ’丨 A Fast Audio Classification from MPEG Coded Data,’’ in Proc. of l CASS?9 9 9 , vo 1 . 6, pp. 3005-3008, Phoenix, USA May 15-19, 1999.[18] Yasuyuki Nakajima, Yang Lu, Masaru Suganno, Akio Yoneyama, Hiromasa Yanagihara, and Akira Kurematsu, '丨 A Fast Audio Classification from MPEG Coded Data,' 'in Proc. Of l CASS? 9 9 9, vo 1 6, pp. 3005-3008, Phoenix, USA May 15-19, 1999.
[1 9 ] J . Nam and A. H. Tewfik,’’Combined audio and visual streams analysis for video sequence segmentation,丨,in Pr oc. of ICASSP, volume 4, pages 2665〜2668, 1997.[1 9] J. Nam and A. H. Tewfik, ’’ Combined audio and visual streams analysis for video sequence segmentation, 丨, in Pr oc. Of ICASSP, volume 4, pages 2665 ~ 2668, 1997.
[2 0 ] W e i Qi,Lie Gu, Hao Jiang,X i ang-Rong Chen and Hong-Jiang Zhang, M Integrating Visual, Audio[2 0] We e Qi, Lie Gu, Hao Jiang, X i ang-Rong Chen and Hong-Jiang Zhang, M Integrating Visual, Audio
第17頁 1223162, 雙正替篸頁 奋年1渺Page 17 1223162, Shuangzheng for the title page Fen Nian 1
E 五、發明說明(12)E V. Description of the invention (12)
And Text Analysis For News Video", in 7th IEEE Intn’ l Conference on Image Processing (ICIP 2000), Vancouver, British Columbia, Canada, 10-13 September 2000 [2l ]A . E. Rosenberg, I . Magrin-Chagnolleau, S.And Text Analysis For News Video ", in 7th IEEE Intn 'l Conference on Image Processing (ICIP 2000), Vancouver, British Columbia, Canada, 10-13 September 2000 [2l] A. E. Rosenberg, I. Magrin-Chagnolleau, S.
Par thasarathy, and Q. Huang,丨,Speaker detection in broadcast speech databases,丨丨 in Pr oc. of International Conference on Spoken Language Processing, Sydney, November l 9 9 8.Par thasarathy, and Q. Huang, 丨, Speaker detection in broadcast speech databases, 丨 丨 in Pr oc. Of International Conference on Spoken Language Processing, Sydney, November l 9 98.
[2 2 ] J i nx i Xu and W. Bruce Croft, ’丨 Improving the effectiveness of information retrieval with local context analysis, f, in ACM t r ans. on information system, vol. 18, no. 1, pp. 79-112, Jan. 2 0 0 0.[2 2] J i nx i Xu and W. Bruce Croft, '丨 Improving the effectiveness of information retrieval with local context analysis, f, in ACM tr ans. On information system, vol. 18, no. 1, pp. 79 -112, Jan. 2 0 0 0.
[2 3 ] H. Wactlar, Hauptmann, A . , Smith, M.,[2 3] H. Wactlar, Hauptmann, A., Smith, M.,
Pendy a 1 a, K . , f, Automated Video Segmentation for On-Demand Retrieval from Very Large Video Libraries,丨’ in the 13th SMPTE (Society of Motion Picture and Television Engineers) Technical Conference and World Expo, Los Angeles, CA, 1 9 9 5.Pendy a 1 a, K., F, Automated Video Segmentation for On-Demand Retrieval from Very Large Video Libraries, 丨 'in the 13th SMPTE (Society of Motion Picture and Television Engineers) Technical Conference and World Expo, Los Angeles, CA, 1 9 9 5.
第18頁Page 18
置 裝 聞 新 體 媒 多 成 轉 一 J 自 聞 新 視 訊 資 鐽 關 及 析 分 容 内 聞 新 。訊 例視 範將 訊, 視明 聞發 新本 視據 電根 是是 6 7 3 8 rv /VV /(\ /IV 圖圖圖圖圖統圖 統 Ο 〇 卩統統 系 - 系系 子Λ U 彳之之 害 c C >識識 r辨辨 元 #音字 $語文 彳播幕 #主螢 ο ο 。 系 0)子 22之 C 聞 統新 系字 子文 之尋 取搜 擷路 面網 畫際 鍵網 關在 將, 明明 發發 本本 據據 根根 是是 產 字 t ^ 鍵 關 及 析 分 聞 新 字 文 頁 網 將 明 發 本 據 根 是 ο 3 /IV 統 子 之 生 之 聯 關 相 聞 新 頁 網 與 聞 新 訊 視 將 明 發 本 據根” 是1 / 2 9)統 C系 圖子 文 路 網 統 際 系 網 子 及 之 訊 放 視 播 9Μ^ Λ Ά 新。新 位0)位 數(6數 將統將 ,系, 明子明 發之發 本庫本 據料據 根資根 是入是 }存} nu 3¾ 11 1i 1i C新c 圖字圖 4 圖 明 發 本 據 根 是 明 發 。本 0)據 5 根 統是 N)/ 子13 之C 訂圖 數 建 將 所 修 及 輯 料 資 頁 網 聞 新 位 編 料 資 之 頁 網 聞 新 位 數 立 a 圖 中 其 内 聞 新 改 修 是 势} 曰 C 擇圖 選·, 是面 介 之 元 單 聞 厂新 51擇 C選 面、巳 介:^ 訂:s 修圖 及; 輯面 介 之 台 ^a 及The installation of news media can be converted into a new one. Self-viewing video information gateway and analysis of content and news. The example is based on the general information, and the new information is based on the fact that the new video is 6 7 3 8 rv / VV / (/ / IV Figures, Figures, Figures, Figures, Figures, and Figures 0 〇 卩 系 系-System Λ U The damage of cc c > Identification r 辨 定 元 元 # 音 字 $ 语文 彳 播 幕 # 主 萤 ο ο. Department 0) C 22 子 C Wen Tongxin Department of Chinese text search search road network The key gateway is about to publish the document, which is based on the word t ^ Key Guan and analysis of the new text page. The web page will publish the document, which is the new page of the relevant issue of the birth of 统 3 / IV Wang and Wenxin News will publish the root cause "is 1/2 9) System C, Tuziwen Road, Network, Intersystem and Network News 9MV ^ Λ Ά New. New 0) Digits (Six generals, general, Mingzi Mingfa's issue of the library, according to the basis of the source is to deposit} nu 3¾ 11 1i 1i C new c Figure Figure 4 Figure Mingfa is based on Mingfa . The 0) According to the 5 system is N) / Sub 13 The C order drawing number will be repaired and edited. Within smell new change repair is a potential} said C Optional FIG selected from *, is the surface referral membered monocyclic smell plant new 51 Optional C selected from the surface, had medium: ^ Order: S retouching and; series surface referral stage ^ a and
第19頁 1223 麼 圖式簡單說明 容之介面 新聞單元 圖(14 ) 介面(4 2 ) 是使用者 介面。 圖(1 5 ) 圖(1 6 ) 圖(1 7 ) 圖(1 8 ) 圖(1 9 ) 詢介面。 圖(20 ) 展示及使 圖(21 ) 内容展示 圖(22 ) 影像展示 ;圖(d )是分割新聞單元之介面;圖(e )是合併 之介面。 是根據本發明,所建立之數位新聞播放使用者 ;其中圖(a)是選擇日期及電台之介面;圖(b) 選擇新聞單元之介面;圖(c )是播放新聞内容之 是根據本發明之使用者登入流程圖。 是根據本發明,所建立之使用者登入介面。 是根據本發明,所建立之使用者註冊介面。 是根據本發明之行動網頁瀏覽與查詢流程圖。 是根據本發明之行動網頁,所建立之使用者查 是根據本發明之行動網頁 用者點選之介面。 是根據本發明之行動網頁 之介面。 是根據本發明之行動網頁 之介面。 所建立之新聞標題 所建立之新聞文字 所建立之新聞關鍵 圖式元件符號說明 10 11 20 類比新聞視訊源 類比數位轉換器 視訊新聞内容分析及關鍵資訊產生子系統Page 19 1223 Mod Schematic description of Rongzhi's interface News unit Figure (14) The interface (4 2) is the user interface. Figure (1 5) Figure (16) Figure (1 7) Figure (1 8) Figure (1 9) Inquiry interface. Figure (20) shows and uses Figure (21) Content display Figure (22) Image display; Figure (d) is the interface for dividing news units; Figure (e) is the interface for merging. It is a digital news broadcast user created according to the present invention; wherein (a) is an interface for selecting a date and a radio station; (b) is an interface for selecting a news unit; and (c) is an interface for broadcasting news content according to the present invention. User login flowchart. It is a user login interface created according to the present invention. It is a user registration interface created according to the present invention. It is a flowchart of mobile web browsing and querying according to the present invention. It is the mobile webpage according to the present invention, and the user search interface created by the user is the interface clicked by the user. Is the interface of the mobile web page according to the present invention. Is the interface of the mobile web page according to the present invention. Created news title Created news text Created news key Graphic component symbol description 10 11 20 Analog news video source Analog digital converter Video news content analysis and key information generation subsystem
第20頁 1 1223162 圖式簡單說明 20 1 數 位 音 訊 及 影 訊 分 割 210 新 聞 單 元 分 割 系 統 220 關 鍵 晝 面 擷 取 之 系 統 230 螢 幕 文 字 辨 識 之 系 統 240 主 播 語 音 辨 識 之 系 統 2410 2 4 2 0 2 4 3 0 2440 2 4 5 0 新聞單元語音資料 語音抽樣及分割 語音特徵抽取 語音辨識器 主播語音資料庫 2 1 視 訊 新 聞 與 網 頁 新 聞 相 關 聯 子 系 統 30 網 頁 文 字 新 聞 分 析 及 關 鍵 字 產 生 子 系 統 31 網 頁 新 聞 文 字 暫 存 資 料 庫 32 網 際 網 路 搜 尋 文 字 新 聞 子 系 統 33 網 頁 新 聞 編 輯 及 分 類 子 系 統 40 數 位 新 聞 播 放 子 系 統 42 數 位 新 聞 播 放 使 用 者 介 面 50 數 位 新 聞 網 頁 資 料 編 輯 及 修 訂 子 系 統 51 數 位 新 聞 網 頁 之 資 料 編 輯 及 修 訂 介 面 52 編 輯 者 60 數 位 新 聞 視 訊 及 網 際 網 路 文 字 新 聞 資 料庫子系統Page 20 1 1223162 Brief description of the diagram 20 1 Digital audio and video segmentation 210 News unit segmentation system 220 Key daytime acquisition system 230 Screen text recognition system 240 Anchor speech recognition system 2410 2 4 2 0 2 4 3 0 2440 2 4 5 0 News unit voice data Voice sampling and segmentation voice feature extraction Voice recognizer anchor voice database 2 1 Video news and web news correlation subsystem 30 Web text news analysis and keyword generation subsystem 31 Web news text temporarily Store database 32 Internet search text news subsystem 33 Web news editing and classification subsystem 40 Digital news playback subsystem 42 Digital news playback user interface 50 Digital news page data editing and revision subsystem 51 Digital news page data editing And revision interface 52 editor 60 digital news video and internet Road text word news database subsystem
第21頁Page 21
Claims (1)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW91105842A TWI223162B (en) | 2002-03-25 | 2002-03-25 | Method and computer system for automatic generation of multimedia WWW news contents from broadcast news video |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW91105842A TWI223162B (en) | 2002-03-25 | 2002-03-25 | Method and computer system for automatic generation of multimedia WWW news contents from broadcast news video |
Publications (1)
Publication Number | Publication Date |
---|---|
TWI223162B true TWI223162B (en) | 2004-11-01 |
Family
ID=34546046
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW91105842A TWI223162B (en) | 2002-03-25 | 2002-03-25 | Method and computer system for automatic generation of multimedia WWW news contents from broadcast news video |
Country Status (1)
Country | Link |
---|---|
TW (1) | TWI223162B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7742111B2 (en) | 2005-05-06 | 2010-06-22 | Mavs Lab. Inc. | Highlight detecting circuit and related method for audio feature-based highlight segment detection |
TWI474200B (en) * | 2012-10-17 | 2015-02-21 | Inst Information Industry | Scene clip playback system, method and recording medium |
TWI493363B (en) * | 2011-12-28 | 2015-07-21 | Intel Corp | Real-time natural language processing of datastreams |
TWI497983B (en) * | 2010-09-29 | 2015-08-21 | Accton Technology Corp | Internet video playback system and its method |
CN113763959A (en) * | 2021-10-19 | 2021-12-07 | 康佳集团股份有限公司 | Voice control method, device, terminal and storage medium based on information reorganization |
TWI784913B (en) * | 2022-05-25 | 2022-11-21 | 中華電信股份有限公司 | A channel program hot spot detection system, method and computer-readable medium thereof |
-
2002
- 2002-03-25 TW TW91105842A patent/TWI223162B/en not_active IP Right Cessation
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7742111B2 (en) | 2005-05-06 | 2010-06-22 | Mavs Lab. Inc. | Highlight detecting circuit and related method for audio feature-based highlight segment detection |
TWI497983B (en) * | 2010-09-29 | 2015-08-21 | Accton Technology Corp | Internet video playback system and its method |
TWI493363B (en) * | 2011-12-28 | 2015-07-21 | Intel Corp | Real-time natural language processing of datastreams |
US9710461B2 (en) | 2011-12-28 | 2017-07-18 | Intel Corporation | Real-time natural language processing of datastreams |
US10366169B2 (en) | 2011-12-28 | 2019-07-30 | Intel Corporation | Real-time natural language processing of datastreams |
TWI474200B (en) * | 2012-10-17 | 2015-02-21 | Inst Information Industry | Scene clip playback system, method and recording medium |
CN113763959A (en) * | 2021-10-19 | 2021-12-07 | 康佳集团股份有限公司 | Voice control method, device, terminal and storage medium based on information reorganization |
CN113763959B (en) * | 2021-10-19 | 2024-01-26 | 康佳集团股份有限公司 | Voice control method, device, terminal and storage medium based on information recombination |
TWI784913B (en) * | 2022-05-25 | 2022-11-21 | 中華電信股份有限公司 | A channel program hot spot detection system, method and computer-readable medium thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9286360B2 (en) | Information processing system, information processing device, information processing method, and computer readable recording medium | |
CN100515078C (en) | Media asset management system for managing video news segments and associated methods | |
US9407942B2 (en) | System and method for indexing and annotation of video content | |
US20110022589A1 (en) | Associating information with media content using objects recognized therein | |
Saba et al. | Analysis of vision based systems to detect real time goal events in soccer videos | |
US20090259623A1 (en) | Systems and Methods for Associating Metadata with Media | |
US20150052128A1 (en) | Query response using media consumption history | |
BR112016006860B1 (en) | APPARATUS AND METHOD FOR CREATING A SINGLE DATA FLOW OF COMBINED INFORMATION FOR RENDERING ON A CUSTOMER COMPUTING DEVICE | |
Nandzik et al. | CONTENTUS—technologies for next generation multimedia libraries: Automatic multimedia processing for semantic search | |
TWI223162B (en) | Method and computer system for automatic generation of multimedia WWW news contents from broadcast news video | |
CN100505072C (en) | Method, system and program product for generating a content-based table of contents | |
Ronfard et al. | A framework for aligning and indexing movies with their script | |
Lian | Innovative Internet video consuming based on media analysis techniques | |
US8264727B2 (en) | Data processing apparatus, method, program, and storage medium for setting identification information based on metadata, and advantageously displaying print data | |
CN114845149A (en) | Editing method of video clip, video recommendation method, device, equipment and medium | |
WO2015094311A1 (en) | Quote and media search method and apparatus | |
Tseng et al. | Video personalization and summarization system | |
JP4755122B2 (en) | Image dictionary generation method, apparatus, and program | |
AT&T | ||
CN113486212A (en) | Search recommendation information generation and display method, device, equipment and storage medium | |
Smith | The search for interoperability | |
Gibbon et al. | Automated content metadata extraction services based on MPEG standards | |
Geisler et al. | Crowdsourcing the indexing of film and television media | |
Rayar et al. | A large-scale TV video and metadata database for French political content analysis and fact-checking | |
Liu et al. | Content personalization and adaptation for three-screen services |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
MM4A | Annulment or lapse of patent due to non-payment of fees |