201026039 九、發明說明： 【發明所屬之技術領域】 本發明相關於一種替照片加入標籤之系統及方法，尤指一種 替照片加入地理標籤及依據照片内容所產生之影像描述標籤之系 統及方法。 ❹ 【先前技術】 近年來，在網路上與人分稍片已逐_為風氣，為了讓他 人依個人興趣迅速且輕易地搜尋到照片，一般會在相片内加入標 籤(Tag)。舉例來說，兩位朋友張三和李四在慶祝張三生日時拍照 留念，則可在相片内加人厂張三」、「李四」和「生日」等標藏f 為了增加個人網路分享空間的閱覽率，越來越多數量的相片被加 入標籤，目前這是網路發展的趨勢。 然而，在相片内加入標籤的過程費時費力，使用者往往傾向 使用最低數量的標籤’如此當然無法達到搜尋照片的最佳效果， 因此需要-種能自動替照片產生標籤之系統和方法。 【發明内容】 本發明相關於-種依據照片内容及地理資訊自動替照片加入 標籤之方法，包含使用-個人導航裝置的照械來拍攝一昭片： 使用該個人導航裝置產生-相關於該照片之地理標籤，並將該地 6 201026039 理標籤加入該照片以產生一定位照片；將該定位照片傳送至一 OCR伺服器；使用該OCR伺服器針對該定位照片執行光學字型 辨識，並依據由該定位照片辨識出之一文字内容產生影像描述標 籤；以及使用該個人導航裝置在該些影像描述標籤中進行選取， 並將選取出之影像描述標籤加入該定位照片。 本發明另相關於一種依據照片内容及地理資訊自動替照片加 ❹入標籤之個人導航裝置，包含一照相機，用來拍攝一照片、產生 一相關於該照片之地理標籤，以及將該地理標籤加入該照片以產 生一定位照片；一通訊埠，用來將該定位照片傳送至一 〇CR伺服 器，以及在該OCR伺服器針對該定位照片執行光學字型辨識並依 據由該定位照片辨識出之一文字内容產生影像描述標籤後，接收 該些影像描述標籤；以及一使用者介面，用來在該些影像描述標 籤中進行選取，並將選取出之影像描述標藏加入該定位照片。 ❹ 【實施方式】 本發明使用光學字型辨識技術(Optical Character Recognition , 0CR)來辨識照片内的文字内容，進而產生有可能被 使用的照片標籤。OCR技術是用來分析數位影像、辨識數位影像 内的文字内容，以及將文字内容以一特定格式輸出，例如以文件 格式。此外，照片亦會加入相關於拍攝地點的地理標藏(Ge〇tag)。 這些標籤可自動產生’使用者不需要人工加入標籤。 201026039 請參考第1圖’第1圖内所顯示之照片10包含了可使用〇CR 技術辨識出之文子内容。照片10的拍攝地點是在一家電腦用品商 店前方，因此可看見店面前張貼的招牌12、14、16、18和2〇。在 第1圖的例子中’招牌12、14、16、18和20包含字母形式或數 字形式舶容，可輕§地透财肖（〕CR麟來賴_容。招牌 12顯示了此電腦用品商店之店名’招牌16顯示了被拍攝商店的類 型’而招牌18和20 ,顯示了此電腦用品商店所提供的服務和產品。 ❹招牌12、14、16、18和20各分別描述了被拍攝商店的某種特性， 因此每一招牌都適合用來當作照片1〇的標籤。 假设將照片ίο上傳至熱門的照片分享網站(如Flickr)，若依據 招牌I2、14、16、18和20產生標籤，使用者則可透過對應於這 些標籤之任一關鍵字來搜尋到照片1〇。此外，透過照片1〇内相關 於拍攝地點的地理標籤，使用者可得知此電腦用品商店的準確位 置。 ❹ 近年來，使用全球定位系統（Gbbal p〇siti〇ning System，Gps彡 之個人導航裝置多半具有内建照相機，已經廣泛地應用於車内導 航系統。個人導航裝置的常見功能包含提供一地圖資料庫以產生 導航指示，並將導航指示顯示於個人導航裝置。在車内應用時， -般會使用吸座或其他方式將個人導航裝置設置於車内儀表板之 上。另一方面’個人導航裝置亦可被㈣地取出車外，在步行時 隨身攜帶。由於個人導航裝置之内建照相機已包含一 Gps接收 8 201026039 器，地理標籤可被加入個人導航裝置所拍攝之照片内。 凊參考第2圖，第2圖說明了本發明中一種能自動替照片產 生標籤之系統100。系統100包含一個人導航裝置110和其内建之 照相機112 ’照相機112可拍攝如第1圖所示之照片1〇。同時， 個人導航裝置110可將地理標籤加入照片1〇，如此就能夠辨識出 照片10的拍攝地點。之後，個人導航裝置11〇將照片10透過一 ❹網際網路140傳送至一閘道器150。 將照片10從個人導航裝置11〇傳送至閘道器150的較佳實施 方式有兩種。在第一種較佳實施方式中，個人導航裝置透過 自身的通訊埠116連接至一電腦13〇之通訊埠132。電腦130可利 用其内安裝之軟體134和個人導航裝置11〇進行通訊，並接收個 人導航裝置110傳來之照片10,接著再將照片1〇透過網際網路 140傳送至閘道器150。 在第二種較佳實施方式中，個人導航裝置110使用無線傳輸 技術（例如IEEE 802.11x標準）將照片1〇傳送至閘道器150。在 此種清瓜下’個人導航裝置Π〇會包含一天線118，因此能以無線 方式收發資料。使用者可透過個人導航裝置之使用者介面114 來選取欲傳送至閘道器15〇之照片1〇。 不論是使用第一或第二種較佳實施方式，當閘道器15〇收到 201026039 照片10後，會將照片10傳送至一 OCR伺服器16〇()〇CR伺服器 160接著針對照片10執行習知〇(：11技術，以辨識出照片1〇内招 牌12、14、16、18和20上的文字内容，最後再將辨識出的文字 内容傳回閘道器15〇。 在前述第一種較佳實施方式中，個人導航裝置11〇是透過電 腦130連接至閘道器150,因此閘道器15〇會將辨識出的文字内容 ❹傳回電腦13〇。此時，使用者可利用電腦130内安裝之軟體134， 針對OCR伺服器160所辨識出的文字内容進行編輯、增加或刪 除。當使用者決定哪些文字内容相關於照片1〇後，編輯後的文字 内容會被儲存並加入照片1〇以做為標籤。此時，照片1〇包含先 前加入的地理標籤，以及依據OCR技術辨識照片1〇之文字内容 後所產生的影像描述標籤。之後，使用者可利用電腦13〇内安裝 之軟體134將加入標籤之照片10透過閘道器15〇上傳至一照片分 ❹享網站Π0。使用者亦可只將加入標籤之照片1〇存入一相片飼服 器或同時上傳至照片分享網站170和存入相片伺服器。 在前述第二種較佳實施方式中，個人導航裝置n〇是無線方 式連結至開道器150,因此閘道器15〇會將辨識出的文字内容直接 透過天線118傳回個人導航裝置11〇。此時，使用者可利用個人導 航裝置110之使用者介面114 ’針對OCR伺服器16〇所辨識出的 文子内容進行編輯、增加或刪除。當使用者決定哪些文字内容相 關於照片10後，編輯後的文字内容會被儲存並加入照片以做 201026039 為“籤。此時，照片10 &含先前加入的地理標藏以及依據〇CR 技術辨識照片1〇之文字内容後所產生的影像描述標藏。之後，使 用者可利用使用者介面114將加人標籤之照片1()透過閘道器15〇 上傳至照片分享網站170。 清參考第3圖’第3圖之流程圖說明了本發明中將標籤加入 照片之方法。在步驟2〇〇起始後，於步驟2〇2中使用個人導航裝 β置110之照相機 112拍攝照片1〇。在步驟204中，個人導航裝置 110將地理標籤加入照片10。在步驟206中，加入地理標藏之照 片10會透過f腦130献直接以無線方式傳送至閘道器15〇。在 步驟208中，閘道器15〇將加入地理標籤之照片1〇傳送至〇cr 伺服器160。在步驟210中，OCR伺服器160針對照片1〇執行 OCR技術以辨識出照片10内的文字内容。在步驟212中，〇cr 伺服器160將辨識出的文字内容傳回電腦13〇或個人導航裝置 珍110。在步驟214中，使用者針對0CR伺服器16〇所辨識出的文 字内容進行編輯、增加或刪除，再將修改後之文字内容儲存並加 入照片10以做為影像描述標籤。在步驟216中，加入前述地理標 籤及新增影像描述標籤之照片10會透過閘道器15〇上傳至照片分 子網站170。最後於步驟218結束本發明中將標籤加入照片之方 法。 除了能利用OCR技術自動產生影像描述標籤，亦可依據照片 10的分析結果產生其他標籤，例如描述照片之主要顏色或纟士構 11 201026039 的仏籤。先别技術之方式可用分析照片IQ以決定照片的美學 特色（如主要顏色），之後加入照片ΐθ的影像描述標藏則可包含 這些美學特色。此外，本產生之標籤也可宣傳特定產品， 例如個人導航裝置110的廠牌和生產廠商聯絡方式。在上傳至照 片分享網站170之前，本發明可依據個人導航裝置11〇之廠牌或 生產個人導航裝置110之廠商的統一資源定位器（Uniform201026039 IX. Description of the Invention: [Technical Field] The present invention relates to a system and method for adding a label to a photo, and more particularly to a system and method for adding a photo tag to a photo and describing the tag based on the image generated by the photo content. ❹ [Prior Art] In recent years, it has been a slogan on the Internet. In order to allow others to quickly and easily search for photos according to their personal interests, tags are usually added to the photos. For example, two friends, Zhang San and Li Si, took photos during the celebration of Zhang San’s birthday. They can add the label “Zhang San”, “Li Si” and “Birthday” in the photo to increase the personal network sharing space. The reading rate, more and more photos have been added to the label, which is currently the trend of the Internet. However, the process of adding tags to a photo is time consuming and laborious, and users tend to use the lowest number of tags. Thus, of course, the best results for searching for photos cannot be achieved, and thus there is a need for a system and method that automatically generates tags for photos. SUMMARY OF THE INVENTION The present invention relates to a method for automatically adding a label to a photo based on photo content and geographic information, including photographing a camera using a personal navigation device: using the personal navigation device to generate - related to the photo a geographic tag, and adding the location 6 201026039 tag to the photo to generate a positioning photo; transmitting the positioning photo to an OCR server; using the OCR server to perform optical font recognition on the positioning photo, and The positioning photo identifies a text content to generate an image description label; and uses the personal navigation device to select among the image description labels, and adds the selected image description label to the positioning photo. The invention further relates to a personal navigation device for automatically adding a label to a photo according to the photo content and the geographic information, comprising a camera for taking a photo, generating a geographical label related to the photo, and adding the geographical label The photo is used to generate a positioning photo; a communication port is used to transmit the positioning photo to a CR server, and the OCR server performs optical font recognition on the positioning photo and is identified according to the positioning photo. After the text content generates the image description label, the image description labels are received; and a user interface is used to select the image description labels, and the selected image description label is added to the positioning photo. ❹ [Embodiment] The present invention uses Optical Character Recognition (OCR) to recognize text content in a photo, thereby generating a photo tag that may be used. OCR technology is used to analyze digital images, recognize text content in digital images, and output text content in a specific format, such as in file format. In addition, the photo will also include the Ge标tag associated with the location. These tags are automatically generated. The user does not need to manually add tags. 201026039 Please refer to Figure 1 The photo 10 shown in Figure 1 contains the text content that can be identified using the 〇CR technology. Photo 10 was taken in front of a computer shop, so you can see the signs 12, 14, 16, 18 and 2 posted in front of the store. In the example of Figure 1, the signboards 12, 14, 16, 18, and 20 contain letters or numbers, which can be used to lighten the money. The store name 'Signboard 16 shows the type of store being photographed' and the signs 18 and 20 show the services and products offered by this computer store. ❹ Signboards 12, 14, 16, 18 and 20 each describe the shot A certain characteristic of the store, so each signboard is suitable for use as a label for photos. Suppose you upload photos ίο to popular photo sharing sites (such as Flickr), depending on the signs I2, 14, 16, 18, and 20 The tag is generated, and the user can search for the photo 1 by any keyword corresponding to the tag. In addition, the user can know the accuracy of the computer product store through the geographical tag related to the shooting location in the photo 1 Location 近年来 In recent years, the use of the Global Positioning System (Gbbal p〇siti〇ning System, Gps彡 personal navigation devices mostly have built-in cameras, has been widely used in car navigation systems. Personal navigation devices often The function includes providing a map database to generate navigation instructions and displaying the navigation instructions on the personal navigation device. When the vehicle is applied, the personal navigation device is generally placed on the in-vehicle dashboard by using a suction cup or other means. In this respect, the personal navigation device can also be taken out of the car and carried around while walking. Since the built-in camera of the personal navigation device already contains a Gps receiving 8 201026039 device, the geotag can be added to the photo taken by the personal navigation device. Referring to Figure 2, a second embodiment of the present invention illustrates a system 100 for automatically generating a photo for a photo. The system 100 includes a human navigation device 110 and a built-in camera 112. The camera 112 can be photographed as shown in Figure 1. At the same time, the personal navigation device 110 can add the geotag to the photo 1 so that the location of the photo 10 can be recognized. Thereafter, the personal navigation device 11 transmits the photo 10 through the Internet 140 to the web site 140. A gateway 150. There are two preferred embodiments for transferring the photo 10 from the personal navigation device 11 to the gateway 150. In a preferred embodiment, the personal navigation device is connected to a communication port 132 of a computer through its own communication port 116. The computer 130 can communicate with the personal navigation device 11 by using the software 134 installed therein, and receive personal navigation. The photo 10 transmitted by the device 110 is then transmitted to the gateway 150 via the Internet 140. In a second preferred embodiment, the personal navigation device 110 uses a wireless transmission technology (e.g., the IEEE 802.11x standard). The photo 1〇 is transmitted to the gateway 150. Under such a melon, the personal navigation device includes an antenna 118, so that the data can be transmitted and received wirelessly. The user can select the photo to be transmitted to the gateway 15 through the user interface 114 of the personal navigation device. Whether the first or second preferred embodiment is used, when the gateway 15 receives the 201026039 photo 10, the photo 10 is transmitted to an OCR server 16(), the CR server 160 is then directed to the photo 10 Execute the syllabus (:11 technique) to identify the text on the signatures 12, 14, 16, 18 and 20 in the photo 1 and finally pass the identified text back to the gateway 15 〇. In a preferred embodiment, the personal navigation device 11 is connected to the gateway device 150 through the computer 130, so that the gateway device 15 transmits the recognized text content back to the computer 13 〇. At this time, the user can The text content recognized by the OCR server 160 is edited, added or deleted by using the software 134 installed in the computer 130. When the user decides which text content is related to the photo, the edited text content is stored and Add a photo as a label. At this time, the photo 1 contains the geotag that was previously added, and the image description label generated after recognizing the text content of the photo according to the OCR technology. After that, the user can use the computer 13〇 Nei An The software 134 uploads the photo 10 of the added tag to the photo sharing website Π0 through the gateway 15 . The user can also save the photo of the tag into the photo feeding device or upload it to the photo sharing at the same time. The website 170 and the photo server are stored. In the second preferred embodiment, the personal navigation device is wirelessly connected to the channel opener 150, so the gateway device 15 transmits the recognized text content directly through the antenna. 118 returns to the personal navigation device 11. At this time, the user can use the user interface 114 of the personal navigation device 110 to edit, add or delete the content of the text recognized by the OCR server 16〇. After the text content is related to the photo 10, the edited text content will be saved and added to the photo to make 201026039 as “sign. At this time, the photo 10 & contains the previously added geographic label and the photo identification according to the 〇CR technology. The image description generated after the text content is marked. After that, the user can use the user interface 114 to upload the tagged photo 1 () to the photo point through the gateway 15 The website 170. The method of adding a label to a photo in the present invention is described with reference to the flowchart of Fig. 3, Fig. 3. After the start of step 2, the personal navigation device is used in step 2〇2. The camera 112 takes a photo. In step 204, the personal navigation device 110 adds a geotag to the photo 10. In step 206, the photo 10 with the geotag is directly transmitted wirelessly to the gateway via the f brain 130 15. In step 208, the gateway 15 transmits the geotagped photo 1 to the 〇cr server 160. In step 210, the OCR server 160 performs OCR technology for the photo 1 to identify the photo 10 The text content inside. In step 212, 〇cr server 160 transmits the recognized text content back to computer 13 or personal navigation device 110. In step 214, the user edits, adds, or deletes the text content recognized by the 0CR server 16〇, and then stores the modified text content and adds the photo 10 as an image description tag. In step 216, the photo 10 incorporating the aforementioned geotag and the new image description tag is uploaded to the photo subsite 170 via the gateway 15. Finally, in step 218, the method of adding a label to a photo in the present invention is ended. In addition to automatically generating image description tags using OCR technology, other tags can be generated based on the analysis results of photo 10, such as describing the main color of the photo or the signature of the gentleman's structure. A technique that can be used to analyze the photo IQ to determine the aesthetic characteristics of the photo (such as the main color), and then add the image description of the photo ΐθ to include these aesthetic features. In addition, the generated label can also promote specific products, such as the label of the personal navigation device 110 and the manufacturer's contact information. Prior to uploading to the photo sharing website 170, the present invention may be based on the label of the personal navigation device 11 or the uniform resource locator of the manufacturer of the personal navigation device 110 (Uniform)
Resource Locator ’ URL)來產生一廣告標籤，並將廣告標籤加入 φ 照片10之中。 此外，除了提供照片10拍攝地點的經緯度外，地理標籤亦可 包含拍攝時間與拍攝日期的資訊。使用者也可將一共通標籤加入 一系列的照片，以標示在拍攝照片10之期間的行程或目的地資 訊。例如，使用者可將「雪山隧道」或「太魯閣國家公園之旅」 的標籤加入一系列的照片。 综上所述，本發明中依據定位和〇CR技術自動產生標籤，因 此能快速且輕易地將標籤加入照片。由於這些加入標籤之照片能 夠在網路上分享，可幫助照片分享網站的發展，同時提高網路搜 尋的效率和準確度。 以上所述僅為本發明之較佳實施例，凡依本發明申請專利範 圍所做之均等變化與修飾，皆應屬本發明之涵蓋範圍。 201026039 - 【圖式簡單說明】 第1圖為一包含可用OCR技術辨識出文字内容之照片。 第2圖為本發明中一種能自動替照片產生標籤之系統的示意圖。 第3圖為本發明中將標籤加入照片之方法的流程圖。 【主要元件符號說明】 10 照片 100 系統 ❹ 110 個人導航裝置 112 照相機 114 使用者介面 116、132 通訊埠 118 天線 130 電腦 134 軟體 140 網際網路 150 閘道器 160 OCR伺服器 170 照片分享網站 200-218 步驟 12、14、 16、18、20 招牌 13Resource Locator 'URL' to generate an ad tag and add the ad tag to φ photo 10. In addition, in addition to providing the latitude and longitude of the location where the photo 10 is taken, the geotag can also contain information on the shooting time and shooting date. The user can also add a common label to a series of photos to indicate the itinerary or destination information during the photo taken. For example, users can add a series of photos to the label "Snow Mountain Tunnel" or "Taro National Park Tour". In summary, in the present invention, the label is automatically generated in accordance with the positioning and 〇CR technology, so that the label can be quickly and easily added to the photo. Since these tagged photos can be shared on the web, they can help the photo sharing site evolve while improving the efficiency and accuracy of web search. The above are only the preferred embodiments of the present invention, and all changes and modifications made to the scope of the present invention should fall within the scope of the present invention. 201026039 - [Simple description of the diagram] Figure 1 is a photo containing the text content that can be identified by OCR technology. Figure 2 is a schematic diagram of a system for automatically generating a label for a photo in the present invention. Figure 3 is a flow chart of a method of adding a label to a photo in the present invention. [Main component symbol description] 10 Photo 100 System ❹ 110 Personal navigation device 112 Camera 114 User interface 116, 132 Communication 埠 118 Antenna 130 Computer 134 Software 140 Internet 150 Gateway 160 OCR Server 170 Photo Sharing Website 200- 218 steps 12, 14, 16, 18, 20 signboard 13