Patent Application
20080097981
Illustrative embodiments of the present invention are described in detail below with reference to the attached drawing figures, which are incorporated by reference herein and wherein:
Referring initially to
The invention may be described in the general context of computer code or machine-useable instructions, including computer-executable instructions such as program modules, being executed by a computer or other machine, such as a personal data assistant or other handheld device. Generally, program modules including routines, programs, objects, components, data structures, etc., refer to code that perform particular tasks or implement particular abstract data types. The invention may be practiced in a variety of system configurations, including hand-held devices, consumer electronics, general-purpose computers, specialty computing devices, servers, etc. The invention may also be practiced in distributed computing environments where tasks are performed by remote-processing devices that are linked through a communications network.
Network environment 100 includes a client 102 coupled to a network 104 via a communication interface. The communication interface may be an interface that can allow the client to be directly connected to any other device or allows the client 102 to be connected to a device over network 104. Network 104 can include, for example, a local area network (LAN), a wide area network (WAN), or the Internet (or the World Wide Web). In an embodiment, the client 102 can be connected to another device via a wireless interface through a wireless network 104.
One or more servers communicate with the client 102 via the network 104 using a protocol such as Hypertext Transfer Protocol (HTTP), a protocol commonly used on the Internet to exchange information. In the illustrated embodiment, a front-end server 106 and a back-end server 108 (e.g., web server or network server) are coupled to the network 104. The client 102 employs the network 104, the front-end server 106 and the back-end server 108 to access Web page data stored, for example, in a central data index (index) 110.
Embodiments of the invention provide searching for relevant data by permitting search results to be displayed to a user 112 in response to a user-specified search request (e.g., a search query). In one embodiment, the user 112 uses the client 102 to input a search request including one or more terms concerning a particular topic of interest for which the user 112 would like to identify relevant electronic documents (e.g., Web pages). For example, the front-end server 106 may be responsive to the client 102 for authenticating the user 112 and redirecting the request from the user 112 to the back-end server 108.
The back-end server 108 may process a submitted query using the index 110. In this manner, the back-end server 108 may retrieve data for electronic documents (i.e., search results) that may be relevant to the user. The index 110 contains information regarding electronic documents such as Web pages available via the Internet. Further, the index 110 may include a variety of other data associated with the electronic documents such as location (e.g., links, or URLs), metatags, text, and document category. In the example of
A search engine application (application) 114 is executed by the back-end server 108 to identify web pages and the like (i.e., electronic documents) in response to the search request received from the client 102. More specifically, the application 114 identifies relevant documents from the index 110 that correspond to the one or more terms included in the search request and selects the most relevant web pages to be displayed to the user 112 via the client 102.
Content manager 204 may be a server such as a workstation running the Microsoft Windows®, MacOS™, Unix, Linux, Xenix, IBM AIX™, Hewlett-Packard UX™, Novell Netware™, Sun Microsystems Solaris™, OS/2™, BeOS™, Mach, Apache, OpenStep™ or other operating system or platform. In an embodiment, content manager 204 may be a search engine comprising of one or more elements 106, 108, 110, 114, and 116 (
Aggregation component 206 can be utilized to crawl web pages in order to aggregate images and text that appears on the same web pages as the images. The text can include, for example, any type of characters or symbols. Once an image and the text corresponding to the image has been identified, the aggregation component can store the images and their corresponding text in database 208. In an embodiment, database 208 is the same as index 110 (
In an embodiment, content manger 204 can receive search requests including one or more search queries for images stored in database 208. A search query can include any text that relates to an image that a requester would like to retrieve. The content manager 204 can identify the text within the search query and can compare the text with text stored in database 208. The content manager 204 can retrieve any images that have associated text that is similar to the text within the search query. Once the images have been retrieved, ranking component 210 can be utilized to rank the retrieved images in an order of relevance to the search query. The ranking component 210 can determine the order of relevance by using one or more ranking factors for determining relevance. The ranking factors can be used to upgrade or downgrade an image's level of relevance to the search query.
Name detector 214 can be utilized to detect whether a search query includes a name of a person in it. In an embodiment, the name detector 214 can be trained to recognize different names by inputting a list of first names and last names from various name books into name detector 214. Face detector 216 may be any conventional face detector that can be utilized to detect faces in the images stored in database 208.
At operation 306, the identified images can be associated with their corresponding identified text within a database. At operation 308, the images within the database are ranked based on one or more ranking factors. An image's ranking can be upgraded or downgraded based on the ranking factors. Below are some ranking factors that can be employed when ranking images:
One ranking factor can based on the number of websites that contain an identical image. With this ranking factor, the invention can determine that images that appear within the web pages of more than one website may be more relevant than images that appear on only one website. As such, an image's ranking can be upgraded depending on the number of websites that the image is located. In an embodiment, the invention can determine if the different websites contain an identical image by evaluating the URL of the image. If the websites point to the same URL of a particular image, then it can be determined that images are identical. In another embodiment, identical images can be determined by calculating a hash of an image. The hash value of one image can be compared to has values of other images, and if any of the hash values are the same, the images are considered to be identical. As described above, the greater the number of websites that contain identical images to the identified image, the higher the identified image is ranked. However, in another embodiment, the greater the number of websites that contain identical images to the identified image, the lower the identified image is ranked.
Another ranking factor can based on the number of websites that contain similar images. For example, an image on one website that is similar to images on other websites may receive a higher ranking depending on the greater the number of other websites that include these similar images. In an embodiment, a first image located on a first web site is similar to a second image on a second website if the second image is a modified version of the first image. For example, the first and second images are considered to be similar if the second image is a resized version (bigger or smaller) than the first image. In another example, a modified version can also include a second image that has been produced by cropping a first image, or can include a second image that has been produced by adding a border around a first image. As described above, the greater the number of websites that contain similar images to a first identified image, the higher the identified image is ranked. However, in another embodiment, the greater the number of websites that contain similar images to the identified image, the lower the identified image is ranked below other images.
Another ranking factor can be based on the size of the images. In an embodiment, the invention can be configured to determine that users are more likely to click on images with the greater number of pixels. As such, images with a greater number of pixels can be ranked higher than images with a lower number of pixels. In another embodiment, images with a greater number of pixels can be ranked lower than images with a lower number of pixels.
Another ranking factor can be based on the number of link relationships an image has with other images. In an embodiment, images with a link relationship can upgrade each other's ranking. Two images can have a link relationship if in response to accessing a first image a user would be presented with a second image. Such a link relationship can be exhibited, for example, when an image and a thumbnail size version of the image are linked together such that accessing one version of the image may lead the user to be presented with the other version of the image. As described above, the greater the number of link relationships an image has, the higher the image is ranked above other images. However, in another embodiment, the greater the number of link relationships an image has with other images, the lower the image is ranked below other images.
Using a link relationship ranking factor, images that are linked together can share each other's associated metadata that can later be used for responding to search queries. For example, suppose a thumbnail image having text nearby is linked to a larger image of the thumbnail wherein the larger image is displayed on a webpage that has no text. Each image could affect, and could possibly raise, each other's ranking. For example, the larger image's size data, a pixel count for example, can also be shared and associated with the thumbnail, and the text near the thumbnail image can also be shared and associated with the larger image. The shared metadata can be associated with each of the linked images within the database 208.
Another ranking feature can be based on the number of times an image is used within a website. For example, if a plurality of web pages within a website link to the same image (the website contains the image on more than web page within the website) or if an image is displayed a plurality of times on a web page within a website, then the invention can be configured to give the image a lower ranking than other images that are not displayed more than once on a website. The image may receive a lower ranking as the invention may determine that the image is part of the graphic design of the website rather than an important image. However, in another embodiment, an image may receive a higher ranking for the greater number of times the image is displayed within a website.
Another ranking factor can be based on detecting images that do not meet or that exceed image feature levels. In an embodiment, image features may include, but are not limited to, number of pixels, aspect ratio, image file size, image entropy, and image gradient. An administrator, for example, can set threshold levels for each of the image features. In an embodiment, an image may be ranked lower if it does not meet or if it exceeds a threshold level for any of the image features. In another embodiment, an image may be ranked higher if it does not meet or if it exceeds a threshold level for any of the image features.
Another ranking feature can be based on the total number of images on a webpage. For example, an image may be given a higher or lower ranking based on the number of images that are on the same web page as the image. Additionally, another ranking feature can be based on the total number of images that are linked to by a particular web page. For example, an image may be given a higher or lower ranking based on the number of images that are linked to by the same web page that the image is located. Moreover, another ranking feature can be based on the total number of thumbnail images that are located on the same webpage as the ranked image. For example, an image may be given a higher or lower ranking based on the number of thumbnail images that are located on the same page as the image. Furthermore, another ranking feature can be based on the total number of links there are to the URL of the an image. For example, an image can be given a higher ranking if it has a greater number of links to its URL compared to other images. In other embodiments, the image may be given a lower ranking if it has a greater number of links to its URL compared to other images.
Another ranking factor can be based on the distance that text within a search query is located on the same web page as an image, such that text that is closer to the image is associated more strongly than text that is further away from the image. The distance that text within a search query is from an image can be based on different distance elements. Distance elements may include the number of intervening words between the text and the image, number of intervening full stops such as “.” “?” “!” and other sentence-ending punctuation/symbols between the text and the image, number of intervening table data tags (<td>) between the text and the image, and the number of intervening table rows tags (<tr>) between the text and the image. An administrator may be able to configure this ranking factor to weigh each of the distance elements equally or differently. For example, the following is an exemplary formula for calculating the distance from an image to text within a search query that is located on the same web page as the image:
Distance=(1*numwords)+(10*numfullstops)+(5*numTDs)−(10*numTRs)
In the above formula, the number of intervening words (numwords) is multiplied by a weight factor of 1, the number of intervening full stops (numfullstops) is multiplied by a weight factor of 10, the number of intervening table data tags (numTDs) is multiplied by a weight factor of 5, and the number of intervening table row tags (numTRs) is multiplied by a weight factor of 10. With this ranking factor and with any of the other ranking factors described above in which a number value is calculated or determined in order to rank an image, the number value can be converted into a score “S” using a sigmoid function to further evaluate the ranking of images. With the present ranking factor, if the search query includes one or more words, the distance for each word can be converted into a score “S” and each of the scores can be summed in order to compute an overall score for the entire search query.
Another ranking factor can be based on whether an image includes a face of a person in it. For example, a conventional face detector 216 (
Additionally, the invention can be configured to evaluate whether the received search query is a request for an image of a person. In an embodiment, the invention can determine whether the search query is a request for an image of a person through use of a name detector 214 (
While particular embodiments of the invention have been illustrated and described in detail herein, it should be understood that various changes and modifications might be made to the invention without departing from the scope and intent of the invention. The embodiments described herein are intended in all respects to be illustrative rather than restrictive. Alternate embodiments will become apparent to those skilled in the art to which the present invention pertains without departing from its scope.
From the foregoing it will be seen that this invention is one well adapted to attain all the ends and objects set forth above, together with other advantages, which are obvious and inherent to the system and method. It will be understood that certain features and sub-combinations are of utility and may be employed without reference to other features and sub-combinations. This is contemplated and within the scope of the appended claims.
