Rule-based scene extraction from video

2003

Abstract. The popularity of digital video is increasing rapidly. To help users navigate libraries of video, algorithms that automatically index video based on content are needed. One approach is to extract text appearing in video, which often reflects a scene's semantic content. This is a difficult problem due to the unconstrained nature of general-purpose video. Text can have arbitrary color, size, and orientation. Backgrounds may be complex and changing.

Proceedings 2003 International Conference on Image Processing (Cat. No.03CH37429), 2003

Knowledge-based fuzzy inference and neural learning are used in this paper in order to model the event recognition task in semantic video analysis. The advantage of their use is the symbolic nature of the representation of the knowledge concerning the events to be recognized. Moreover, this knowledge can be adapted with the aid of data taken from video sequences. The proposed system has been tested in soccer video sequences for detecting some complex predetermined (and represented in the form of rules) events.

2001

Lecture Notes in Computer Science, 2000

In this paper, we propose a new approach for high level segmentation of a video clip into shots using spatio-temporal relationships between objects in video frames. The technique we use is simple, yet novel and powerful in terms of effectiveness and user query satisfaction. Video clips are segmented into shots whenever the current set of relations between objects changes and the video frames where these changes have occurred are chosen as key frames. The topological and directional relations used for shots are those of the key frames that have been selected to represent shots and this information is kept, along with key frame intervals, in a knowledge-base as Prolog facts. We also have a comprehensive set of inference rules in order to reduce the number of facts stored in our knowledge-base because a considerable number of facts, which otherwise would have to be stored explicitly, can be derived by these rules with some extra effort.

International Journal of Computer Applications, 2017

Now a day's increase in access of video-based application has opened need for extracting the content in videos. Unprocessed data and low-level features alone are not sufficient to complete the user's need so deeper understanding of the content at the semantic level is required. Currently, manual techniques which are inefficient, subjective and expensive in time and limit the querying capabilities are used to fulfill the gap between low-level representative features and high-level semantic content. The system that allows the user to query and retrieve associated objects, events, and concepts automatically is proposed .The events can also be representative objects, actions, their impressions, and so on. Here an ontology-based video semantic content model which uses spatial/temporal relations in event and concept definitions is leveraged. Simple & efficient process consideration on main object detection & its common associated mixed direct measurable feature like shape, texture & derived features like co-occurrence & topology is evaluated. An ontology definition provides a wide-domain applicable rule construction standard. In addition to domain ontology's, additional rule definitions to lower spatial relation computation cost are used. This leads to describe some complicated

IEEE Transactions on Knowledge and Data Engineering, 2013

Traditional approaches in Multimedia systems force the user to segment video material into simple descriptions, which usually include endpoints and a brief text description in the form of keywords. We propose to segment contextual information into chunks rather than segmenting contiguous frames. The computer can help us organize sets of descriptions that are related to the recorded moving image: it can help us remember what we have shot. Such descriptions can overlap, be contained in, and even encompass a multitude of other descriptions. Parsing moving image sequences is reduced to simply parsing the contextual information which forms the description. Our approach, Stratification, also has important ramifications in terms of an elastic representation of moving images. Ambient sound can provide us with important contextual clues as to what is going on within the frames. Using audio to find patterns of content is an important step towards the eventual automatization of the logging pro...

IEEE Transactions on Circuits and Systems for Video Technology, 2004

A generic system for automatic annotation of videos is introduced. The proposed approach is based on the premise that the rules needed to infer a set of high-level concepts from low-level descriptors cannot be defined a priori. Rather, knowledge embedded in the database and interaction with an expert user is exploited to enable system learning. Underpinning the system at the implementation level is preannotated data that dynamically creates signification links between a set of low-level features extracted directly from the video dataset and high-level semantic concepts defined in the lexicon. The lexicon may consist of words, icons or any set of symbols that convey the meaning to the user. Thus, the lexicon is contingent on the user, application, time and the entire context of the annotation process. The main system modules use fuzzy logic and rule mining techniques to approximate human-like reasoning. A rule-knowledge base is created on a small sample selected by the expert user during the learning phase.

This paper is about extracting knowledge from high-dimensional sets of videos, with a par-ticular reference to the video-surveillance application domain. In this setting "knowledge" refers primarily to the understanding of what are common behaviors in the observed sce-nario, and this will be the main goal of our work. Starting off from data acquired by a still camera, we exploit a behavior modeling pipeline composed of a data-driven abstraction level based on strings and on an unsupervised learning model based on spectral clustering. The estimated patterns of common behaviors are used to associate new temporal data and, consequently, highlight abnormal (or unusual) dynamic events. Experiments on a test set with a ground truth, as well as a set of unlabeled data, allow us to discuss the complexity of selecting appropriate models in the unsupervised domain. We present very good results on a high dimensional set of data acquired from a real video-surveillance system and compa...

Proceedings IEEE International Conference on Multimedia Computing and Systems, 1999

Determining automatically what constitutes a scene in a video is a challenging task, particularly since there is no precise definition of the term "scene". It is left to the individual to set attributes shared by consecutive shots which group them into scenes. Certain basic attributes such as dialogs, settings and continuing sounds are consistent indicators. We have therefore developed a scheme for identifying scenes which clusters shots according to detected dialogs, settings and similar audio. Results from experiments show automatic identification of these types of scenes to be reliable.