A Proposed Framework for Knowledge Extraction from Digital Videos

Abstract

Video contains a huge amount of data which include complex interaction between its elements. Using manual techniques to get video content description is complex process, time consuming and have a lot of limitation as a result of wrong understanding therefore, this paper presents a proposed framework for knowledge extraction from digital video. The framework consists of two phases, the first phase deals with the audio channel and the second phase deals with the video channel. The audio channel passes through three components namely speech recognition, sentence boundary detection and summarization consequently. The second phase also passes through three components namely video segmentation, feature extraction and key frame extraction. The proposed system was applied on online course from Lynda.com (excel data mining fundamentals).The evaluation of the proposed framework is compared with other systems and shows that the proposed framework is efficient. Keywords: speech recognition, sentence boundary detection, summarization, machine learning, pause duration