IJENS Publisher Indexed In SCOPUS


Title: Fused Image Based Video Search Engine
Author(s): Sheena Gupta, R. K. Kulkarni
Pages: 1-6 Paper ID:174305-2929-IJVIPNS-IJENS Published: October, 2017
Abstract: A video search engine is a web-based search engine which crawls the web to find videos stored on digital devices. Video search engine has been given very little attention until the popularity of digital capturing devices and communication tools. Here a novel method is proposed for video search engine using fused image technique. To begin with, the video sequence is split into key-frames which is then used to create the fused image for the sample video. The color and edge histogram features are then extracted from this fused image and stored as a complete feature vector of that respective video. Then on the basis of the query clip, the top five videos are retrieved from the database using Eucilidean distance as a similarity measure. This experiment is performed on a database of 40 video samples spread across 4 categories. Performance evaluation is done with the help of average precision and average recall. The proposed method is also compared with existing video retrieval techniques. Results indicate that proposed algorithm is good compared to existing techniques.
Keywords: Video retrieval, content based video retrieval (CBVR), key-frames, fused image, color moments, edge histogram, query clip.
Full Text (.pdf)  International Journals Of Engineering and Science | 694 KB