Improvement of Telop Recognition Quality by Integrating Web Search Results
نویسندگان
چکیده
Many scenes in recent TV programs display rich text information on the screen as Television Opaque Projectors or telops. Such telops are useful for obtaining information about the scene, or in searching for scenes using keywords. However, the quality of recognized results for telops, especially in Japanese or Chinese, is still poor even using the latest image recognition techniques. In this paper, we propose a method of integrating web search into the recognition process, to improve the quality of telop recognition results, focusing on Japanese news TV programs. At first, the proposed method recognizes telops in the news scenes by a current image recognition method. Next, it searches news web sites for related articles based on search keywords derived from the intermediate image recognition results. Then, the searched articles are used to build a context-based dictionary for correcting errors in recognized characters. We evaluate the proposed method with actual news TV programs and news web sites. The experimental results demonstrate that retrieved web articles are sufficiently related and effective for correcting the incorrectly recognized characters, even though these search keywords are derived from poor quality image recognition results. It indicates that the integration approach is effective in improving the precision of telop recognition results which is applicable for providing searched or summarized TV scenes along with web text as integrated information.
منابع مشابه
Integrating WWW Caches and Search Engines
In this paper we propose the concept of cache plugins, which are customized programs that run WWW cache servers and perform some of the search engine tasks. We describe a prototype implementation of cache plugin to answer client requests directed to a large search engine, using a nearby cache server to store static objects. Experimental results using actual logs show a signiicant improvement on...
متن کاملA New Hybrid Method for Web Pages Ranking in Search Engines
There are many algorithms for optimizing the search engine results, ranking takes place according to one or more parameters such as; Backward Links, Forward Links, Content, click through rate and etc. The quality and performance of these algorithms depend on the listed parameters. The ranking is one of the most important components of the search engine that represents the degree of the vitality...
متن کاملRobust Telop Character Recognition in Video for Content-based Retrieval
The recognition of telop characters in video has two problems: edge degradation and background noise. To overcome these problems, this paper proposes; (1) a feature that describes the shape of background region in addition to character region, (2) a classifier that dynamically decreases the influence of noise using the difference in number of pixel within a local area. Experiments show that the...
متن کاملA Technique for Improving Web Mining using Enhanced Genetic Algorithm
World Wide Web is growing at a very fast pace and makes a lot of information available to the public. Search engines used conventional methods to retrieve information on the Web; however, the search results of these engines are still able to be refined and their accuracy is not high enough. One of the methods for web mining is evolutionary algorithms which search according to the user interests...
متن کاملTowards Supporting Exploratory Search over the Arabic Web Content: The Case of ArabXplore
Due to the huge amount of data published on the Web, the Web search process has become more difficult, and it is sometimes hard to get the expected results, especially when the users are less certain about their information needs. Several efforts have been proposed to support exploratory search on the web by using query expansion, faceted search, or supplementary information extracted from exte...
متن کامل