تصنيف الفيديو باستخذام التنقيب عن البيانات Video Categorization Using Data Mining

Date
2014-04-24
Authors
نور حسين محمود الصليبي
Nour Hussein Mahmoud Alsalieby
Journal Title
Journal ISSN
Volume Title
Publisher
AL-Quds University
جامعة القدس
Abstract
Video categorization using data mining is the area of the research that aims to propose adeveloped method based on Artificial Neural Network (ANN), which could be used to classify video files into different categories according to the content. In order to test this method, the classifications of video files are discussed. The applied system proposes that the video could be categorized in two classes. The first one is educational while is noneducational. The classification is conducted based on the motion using optical flow. Several experiments were conducted using Artificial Neural Network (ANN) model. The research facilitate access to the required educational video to the learners students, especially novice students. This research objective is to investigate how the effect of motion feature can be useful in such lassification. We believe that other effects such audio features, text features, and other factors can enhance accuracy, but this requires wider studies and need more time. The accuracy of results in video classification to educational and non-educational through technique 3 fold cross validation and using (ANN) model is 54%. This result may can be improved by introducing other factors mentioned above.
Description
Keywords
علم الحاسوب, Computer Science
Citation
Collections