Research On 3D Format Detection And Video Clustering Of 3D Video Search Engine

3D movies reach a huge commercial success and bring a rapid development to 3D film industry in recent years. More and more multimedia services are accessible on the Internet due to the development of the Internet and computer performance. All of these help the video broadcast on the Internet to get popular, and these services are accepted and enjoyed by consumers quickly. A large number of 3D devices, such as 3D TV, 3D mobile and 3D game console, are being placed to market as 3D display technology matures. Consumers can enjoy the 3D video with their 3D terminal device even at home now. How to find 3D videos from mass of videos on the Internet is a difficulty for both consumers and video providers.For 3D video searching, full-text search may perform badly because it can not identify the video subject that lots of its search result would have nothing to do with videos. Video search engine is a better choice which is based on full-text search engine and the video meta data extracted from the video page text. But it just limits search subject to video and might return the users some normal videos with 3D relative titles.This thesis analyzes the key technology, content-based 3D identification, of 3D video search engine. The method to improve the user experience, video search results clustering, is also mentioned. In the last section, a design of the implementation of 3D video search engine is proposed, and some detail components are implemented. By analyzing the different implementations of 3D video, different image organization of 3D video, a content-based 3D video identification scheme is summed up.By analyzing the video structure, a video clustering method is achieved for 3D video search engine.