Using Local Region Semantics for Concept Detection in Video.

Published 1 January 2006Open access

Evaggelos Spyrou, George Koumoulos, Yannis Avrithis, Stefanos Kollias

Citations2

TL;DR

This paper presents a framework for the detection of semantic features in video sequences by performing low-level feature extraction on the keyframes of the shots and a “feature vector” including color and texture features is formed.

Abstract

Abstract — This paper presents a framework for the detection of semantic features in video sequences. Low-level feature extraction is performed on the keyframes of the shots and a “feature vector ” including color and texture features is formed. A region “thesaurus ” that contains all the high-level features is constructed using a subtractive clustering method.Then, a “model vector” that contains the distances from each region type is formed and a SVM detector is trained for each semantic concept. Experiments were performed using TRECVID 2005 development data.

Keywords

Computer Science

TechnometricsStatistical Learning Theory

26,913 Citations1999Yuhai Wu, Vladimir Vapnik

Presenting a method for determining the necessary and sufficient conditions for consistency of learning process, the author covers function estimates from small data pools, applying these estimations to real-life problems, and much more.

IEEE Transactions on Circuits and Systems for Video TechnologyColor and texture descriptors

1,650 Citations2001B.S. Manjunath, J.-R. Ohm +2 more

An overview of color and texture descriptors that have been approved for the Final Committee Draft of the MPEG-7 standard is presented, explained in detail by their semantics, extraction and usage.

IEEE Transactions on Neural NetworksSupport vector machines for histogram-based image classification

1,457 Citations1999Olivier Chapelle, Patrick Haffner +1 more

It is observed that a simple remapping of the input x(i)-->x(i)(a) improves the performance of linear SVM's to such an extend that it makes them, for this problem, a valid alternative to RBF kernels.

Extracting Fuzzy Rules from Data for Function Approximation and Pattern Classification

133 Citations2000S. Chiu

Methods for extracting fuzzy rules for both function approximation and pattern classification are presented, based on estimating clusters in the data, which corresponds to a fuzzy rule that relates a region in the input space to an output region.

Lecture notes in computer scienceFusing MPEG-7 Visual Descriptors for Image Classification

63 Citations2005Evaggelos Spyrou, Hervé Le Borgne +4 more

Three content-based image classification techniques based on fusing various low-level MPEG-7 visual descriptors based on a "merging" fusion combined with an SVM classifier, a back-propagation fusion Combined with a KNN classifier and a Fuzzy-ART neurofuzzy network are proposed.

View PDF

Lecture notes in computer scienceLearned Lexicon-Driven Interactive Video Retrieval

39 Citations2006Cees G. M. Snoek, Marcel Worring +2 more

This paper combines automatic learning of a large lexicon of semantic concepts with traditional video retrieval methods into a novel approach to narrow the semantic gap.

View PDF

Kluwer Academic Publishers eBooksIMAGE CLASSIFIERS FOR SCENE ANALYSIS

26 Citations2006Bertrand Le Saux, Giuseppe Amato

An image representation appropriate for scene description is proposed: images are segmented into re- gions and indexed according to the presence of given region types and a classification scheme designed to separate images in the descriptor space is propound.

Region-based video content indexing and retrieval

22 Citations2005Fabrice Souvannavong

This paper shows that LSA performs better on the task of object retrieval despite the quantification process implied, and compares performances of the system with a method using the Earth Mover’s Distance that have the property to keep the original features describing regions.

Time Interval based Modelling and Classification of Events in Soccer Video

7 Citations2003Cees G. M. Snoek, Marcel Worring

This paper proposes to model events in multimodal video by means of time interval relations, to tackle aforementioned problems, and exploits the powerful properties of statistical classifiers, like the Maximum Entropy and Support Vector Machine classifiers.