Development of efficient video compression techniques using semantic based resurgence procedures

Abstract

Nowadays, video compression plays an important role in the day to day life. People use video compression for recreating the video without affecting the quality with reduced size and video transmission area endures several failures owing to the limited amount of a cutting edge technique to store large sized videos. For this reason, video compression method is used. Extracting the semantic data from the large amount of video based applications is the necessary feature. The existing methodologies are insufficient and need high amount of optimization cost. The videos being searchable on the web also have drastically increased. In the first research work, the gap between low-level representative feature and high level semantic content is used. For the deeper understanding, apart from the raw data and low-level features, the content at semantic level is required. The existing periodicity mining was based on audio retrieval which could not be applied to the video. The speedy expansion in the available amount of video data has increased a necessary constraint to lengthen intellectual methodologies to model and extract the semantic content. Characteristic applications in which modelling and extracting video content are decisive include several real-time application. The eventual purpose is to facilitate users to safeguard some desired content from enormous amounts of video data in a competent and semantically consequential approach. The ontology and rule-based model is developed for extracting the semantic contents in videos and is used to track the direction of the spatial movements between two image frames. Ontology provides a vast domain applicable rule production set that allows the user to construct the domain. Rule based model use the spatial/temporal relations to extract the definitions newline