Author(s): S. Gomathy | K.P. Deepa | T. Revathi | L. Maria Michael Visuwasam
Journal: The SIJ Transactions on Computer Science Engineering & its Applications
ISSN 2321-2373
Volume: 01;
Issue: 01;
Start page: 06;
Date: 2013;
VIEW PDF
DOWNLOAD PDF 
Original page
Keywords: Frame Matching | Genre Specific Classification | Image based Search | Multimodality Web Categorization | Multiple Frame Detection | Semantic Indexing and Video Retrieval
ABSTRACT
Searching has become an integral part in today’s Internet bound era. Now-a-days, searching a video content from a large scale video set is difficult because the volume of video increases rapidly that too with the lack of proper tools to handle. Large video collections such as YouTube contains many different genres that searches through Tree- Based Concept and retrieves video result through filename that are subjective and noisy and in many cases not reflecting the accurate content. The overall aim of the paper is to provide ease to the user by giving a refined video search by accepting both text and image inputs. Our system describes architecture for a new video searching mechanism that not only accepts text based inputs but also accepts image based input from the user to retrieve the video results. Here we propose, two step frame work comprising of two levels, genre level and semantic concept level, with a filtering mechanism to generate the accurate video result.
Journal: The SIJ Transactions on Computer Science Engineering & its Applications
ISSN 2321-2373
Volume: 01;
Issue: 01;
Start page: 06;
Date: 2013;
VIEW PDF
DOWNLOAD PDF Original pageKeywords: Frame Matching | Genre Specific Classification | Image based Search | Multimodality Web Categorization | Multiple Frame Detection | Semantic Indexing and Video Retrieval
ABSTRACT
Searching has become an integral part in today’s Internet bound era. Now-a-days, searching a video content from a large scale video set is difficult because the volume of video increases rapidly that too with the lack of proper tools to handle. Large video collections such as YouTube contains many different genres that searches through Tree- Based Concept and retrieves video result through filename that are subjective and noisy and in many cases not reflecting the accurate content. The overall aim of the paper is to provide ease to the user by giving a refined video search by accepting both text and image inputs. Our system describes architecture for a new video searching mechanism that not only accepts text based inputs but also accepts image based input from the user to retrieve the video results. Here we propose, two step frame work comprising of two levels, genre level and semantic concept level, with a filtering mechanism to generate the accurate video result.
