This paper provides an overview of different video content modeling techniques employed in existi... more This paper provides an overview of different video content modeling techniques employed in existing content-based video indexing and retrieval (CBVIR) systems. Based on the modeling requirements of a hypothetical (somewhat ideal) CBVIR system, we analyze and categorize existing modeling approaches. Starting with a review of techniques to model raw video data, we study approaches used to describe physical objects, and conclude with a review on high-level semantic modeling of data with focus on the multimodal analysis. Based on the current status of research in CBVIR systems, we identify the growth potential, future directions, and open research issues. Finally, a hypothetical CBVIR system is outlined in the concluding remarks, which exploits object-based representation of MPEG-4 compressed bitstream and uses multimodal features based on high-level description of video.
We present a novel motion trajectory based video retrieval system using parallel adaptive self or... more We present a novel motion trajectory based video retrieval system using parallel adaptive self organizing maps (PASOMs) in this paper. The motion trajectories are extracted from video sequences by a robust tracker using detection-based particle filter. To reduce the high dimension of motion trajectories and achieve flexible video retrieval, our algorithm first decomposes each raw motion trajectory into subtrajectories by using a maximum acceleration based approach. Each subtrajectory is then modeled and coded by two different methods, polynomial curving fitting and independent component analysis. To fuse the different features of subtrajectories for more efficient and flexible retrieval, we use a PASOMs network as the searching tool where each SOM learns and classifies one feature. Experimental results show the superior performance of the proposed approach for video retrieval comparing with prior approaches.
2003-04-12). http://multimedia. eecs. uic. edu/fai- …, Jan 1, 2003
This paper explores the domain of automatically extracting semantic information from given databa... more This paper explores the domain of automatically extracting semantic information from given database schema and structured textual data for the purpose of Ontology construction. We examine the advantages of ontology based information search mechanisms and survey the difference techniques currently employed for automated ontology construction from given text corpus and from unstructured multimedia data, such as images, audio and video. Finally we propose an algorithm for ontology construction from given structured textual data with the help of database schema. The approach that we thoroughly discuss is from the domain of Formal Concept Analysis which is based on Lattice Theory. We use a generic lexical ontology WordNet to generate word senses and to label the nodes of ontology.
… , 2004. Proceedings of INMIC 2004. 8th …, Jan 1, 2004
In this paper, we outline a novel technique fbr robust indexing and refrieval of video object frn... more In this paper, we outline a novel technique fbr robust indexing and refrieval of video object frnjectoi-ies. Our approach is based on using sputiaternporul crrrilature to represent the trajectories. Daminai7t inflection poirits ,from the czrrvotrr?'e ure ex/racred at mirltiple levels of scale simultaneously reveulii.rg sinrctrtre ai vnwing levels o j details. These inyeclion poiiils, the maxima of Curvattire Scale Space (CSS}, are then used Jbr indexing urd retrieval. E.rpeuirnents 017 a dalahase containing trajectories of high jumps crnd other motions show the ufiirieinvarionl nolure of o w approach.
This figure compares our system in terms of stand-off distance and capture volume with others in ... more This figure compares our system in terms of stand-off distance and capture volume with others in recent literature.
This paper presents novel tensor-based object trajectory modelling techniques for simultaneous re... more This paper presents novel tensor-based object trajectory modelling techniques for simultaneous representation of multiple objects motion trajectories in a content based indexing and retrieval framework. Three different tensor decomposition techniques-PARAFAC, HOSVD and Multiple-SVD-are explored to achieve this goal with the aim of using a minimum set of coefficients and data-dependant bases. These tensor decompositions have been applied to represent full as well as segmented trajectories. Our simulation results show that the PARAFAC-based representation provides higher compression ratio, superior precision-recall metrics, and smaller query processing time compared to the other tensor-based approaches.
Multimedia data, such as text, audio, images and video, is rapidly evolving as the main form for ... more Multimedia data, such as text, audio, images and video, is rapidly evolving as the main form for the creation, exchange, and storage of information in the modern era. Primarily, this is attributed to rapid advances in the three major technologies that determine its growth: VLSI technology that is producing greater processing power; broad-band networks (ISDN, ATM, etc) that are providing much higher bandwidth for many practical applications, and multimedia compression standards (JPEG, H.263, MPEG, MP3, etc) that enable efficient storage and communication. The combination of these three advances is spurring the creation and processing of increasingly high-volume multimedia data, along with its efficient compression and transmission over high-bandwidth networks. This current trend towards the removal of any conceivable bottleneck in using multimedia and its impact on a whole spectrum of users, from advanced research organizations to home users, has led to the explosive growth of visual information available in the form of digital libraries and online multimedia archives. According to a press release by Google Inc. in December 2001, the search engine offers access to over 3 billion web documents and its Image search comprises more that 330 million images. AltaVista has been serving around 25 million search queries per day in more than 25 languages, with its multimedia search featuring over 45 million images, videos and audio clips.
International Workshop on Frontiers of …, Jan 1, 2003
This paper provides an overview of different video content modeling techniques employed in existi... more This paper provides an overview of different video content modeling techniques employed in existing content-based video indexing and retrieval (CBVIR) systems. Based on the modeling requirements of a hypothetical (somewhat ideal) CBVIR system, we analyze and categorize existing modeling approaches. Starting with a review of techniques to model raw video data, we study approaches used to describe physical objects, and conclude with a review on high-level semantic modeling of data with focus on the multimodal analysis. Based on the current status of research in CBVIR systems, we identify the growth potential, future directions, and open research issues. Finally, a hypothetical CBVIR system is outlined in the concluding remarks, which exploits object-based representation of MPEG-4 compressed bitstream and uses multimodal features based on high-level description of video.
We present a compressed domain video object segmentation method for the MPEG encoded video sequen... more We present a compressed domain video object segmentation method for the MPEG encoded video sequences. For a fraction of the raw domain analysis, compressed domain segmentation provides the essential priori information to many vision tasks from surveillance to transcoding that require fast processing of large volumes of data where pixel-resolution boundary extraction is not required. Our method generates accurate segmentation maps in block-resolution at hierarchically varying object levels, which empowers application to determine the most pertinent partition of images. It exploits the block-structure of the compressed video to minimize the amount of data to be processed. All the available motion flow within a group-of-picture are projected onto a single layer that also consists of the frequency decomposition of color pattern. Then, by starting from the blocks where the spatial energy is small, it expands homogeneous regions while automatically adapting local similarity criteria. We also formulate an alternative solution that applies a kernel based clustering where separate spatial, transform, and motion kernels are used to establish the affinity. We show that both region expansion and mean-shift produce similar results as the computationally expensive raw domain segmentation. Finally, a binary clustering iteratively merges the most similar regions to generate a hierarchical partition tree.
Visual Communications and Image Processing, Jan 1, 2007
This paper addresses the issue of multi-source collaborative object tracking in high-definition (... more This paper addresses the issue of multi-source collaborative object tracking in high-definition (HD) video sequences. Specifically, we propose a new joint tracking paradigm for the multiple stream electronic pan-tilt-zoom (EPTZ) cameras. These cameras are capable of transmitting a low resolution thumbnail (LRT) image of the whole field of view as well as a high-resolution cropped (HRC) image for the target region. We exploit this functionality to perform joint tracking in both low resolution image of the whole field of view as well as high resolution image of the moving target. Our system detects objects of interest in the LRT image by background subtraction and tracks them using iterative coupled refinement in both LRT and HRC images. We compared the performance of our joint tracking system with that of tracking only in the HD mode. The results of our experiments show improved performance in terms of higher frame rates and better localization.
Motion information is regarded as one of the most important cues for developing semantics in vide... more Motion information is regarded as one of the most important cues for developing semantics in video data. Yet it is extremely challenging to build semantics of video clips particularly when it involves interactive motion of multiple objects. Most of the existing research has focused on capturing and modelling the motion of each object individually thus loosing interaction information. Such approaches yield low precisionrecall ratios and limited indexing and retrieval performances. This paper presents a novel framework for compact representation of multi-object motion trajectories. Three efficient multitrajectory indexing and retrieval algorithms based on multilinear algebraic representations are proposed. These include: (i) geometrical multiple-trajectory indexing and retrieval (GMIR), (ii) unfolded multiple-trajectory indexing and retrieval (UMIR), and (iii) concentrated multiple-trajectory indexing and retrieval (CMIR). The proposed tensor-based representations not only remarkably reduce the dimensionality of the indexing space but also enable the realization of fast retrieval systems. The proposed algorithms have been implemented and evaluated using real video datasets. Simulation results on both full and partial (segmented) multiple motion trajectories with varying number of objects, trajectory lengths, and sampling rates demonstrate that the CMIR algorithm provides superior precision-recall metrics, and smaller query processing time compared to the other approaches.
In recent years the human iris has established itself as a robust, highly accurate biometric. Aca... more In recent years the human iris has established itself as a robust, highly accurate biometric. Academic research of iris biometrics is expanding and several iris recognition products have become commercially available. Conventional iris recognition systems impose constraints with regard to a ...
In this paper, we propose a novel technique for modelbased recognition of complex object motion t... more In this paper, we propose a novel technique for modelbased recognition of complex object motion trajectories using Gaussian Mixture Models (GMM). We build our models on Principal Component Analysis (PCA)-based representation of trajectories after segmenting them into small units of perceptually similar pieces of motions. These subtrajectories are then fitted with automaticallylearnt mixture of Gaussians to estimate the underlying class probability distribution. Experiments are performed on two data sets; the ASL data set (from UCI's KDD archives) consists of 207 trajectories depicting signs for three words, from Australian Sign Language (ASL); the HJSL data set contains 108 trajectories from sports videos. Our experiments yield an accuracy of 85+% performing much better than existing approaches.
Motion trajectories provide rich spatio-temporal information about an object's activity. The traj... more Motion trajectories provide rich spatio-temporal information about an object's activity. The trajectory information can be obtained using a tracking algorithm on data streams available from a range of devices including motion sensors, video cameras, haptic devices, etc. Developing view-invariant activity recognition algorithms based on this high dimensionality cue is an extremely challenging task. This paper presents efficient activity recognition algorithms using novel view-invariant representation of trajectories thus will recognize motion activities independent of the view point. Towards this end, we derive two Affine-invariant representations for motion trajectories based on curvature scale space (CSS) and centroid distance function (CDF). The properties of these schemes facilitate the design of efficient recognition algorithms based on hidden Markov models (HMMs). In the CSS-based representation, maxima of curvature zero crossings at increasing levels of smoothness are extracted to mark the location and extent of concavities in the curvature. The sequences of these CSS maxima are then modeled by continuous density (HMMs). For the case of CDF, we first segment the trajectory into subtrajectories using CDF-based representation. These subtrajectories are then represented by their Principal Component Analysis (PCA) coefficients. The sequences of these PCA coefficients from subtrajectories are then modeled by continuous density hidden Markov models (HMMs). Different classes of object motions are modeled by one Continuous HMM per class where state PDFs are represented by GMMs. Experiments using a database of around 1750 complex trajectories (obtained from UCI-KDD data archives) subdivided into 5 different classes are reported.
Abstract This paper studies efficient feature spaces for content based indexing and retrieval of ... more Abstract This paper studies efficient feature spaces for content based indexing and retrieval of object motion trajectories. Taking object trajectory data as input, we first investigate highly compact affine invariant feature spaces based on Fourier Descriptors (FD) and Principal Component Analysis (PCA) techniques. Based on these feature spaces, we then develop a hybrid content based indexing and retrieval system that employs a two-stage matching scheme. The first stage uses affine-invariant Fourier Descriptor (FD) for indexing and ...
In this paper, we propose a novel technique for modelbased recognition of complex object motion t... more In this paper, we propose a novel technique for modelbased recognition of complex object motion trajectories using Hidden Markov Models (HMM). We build our models on Principal Component Analysis (PCA)-based representation of trajectories after segmenting them into small units of perceptually similar pieces of motions. These subtrajectories are then grouped using spectral clustering to decide on the number of states for each HMM representing a class of object motion. The hidden states of the HMMs are represented by Gaussian Mixtures (GM's). This way the HMM topology as well as the parameters are automatically derived from the training data in a fully unsupervised sense. Experiments are performed on two data sets; the ASL data set (from UCI's KDD archives) consists of 207 trajectories depicting signs for three words, from Australian Sign Language (ASL); the HJSL data set contains 108 trajectories from sports videos. Our experiments yield an accuracy of 90+% performing much better than existing approaches.
This paper provides an overview of different video content modeling techniques employed in existi... more This paper provides an overview of different video content modeling techniques employed in existing content-based video indexing and retrieval (CBVIR) systems. Based on the modeling requirements of a hypothetical (somewhat ideal) CBVIR system, we analyze and categorize existing modeling approaches. Starting with a review of techniques to model raw video data, we study approaches used to describe physical objects, and conclude with a review on high-level semantic modeling of data with focus on the multimodal analysis. Based on the current status of research in CBVIR systems, we identify the growth potential, future directions, and open research issues. Finally, a hypothetical CBVIR system is outlined in the concluding remarks, which exploits object-based representation of MPEG-4 compressed bitstream and uses multimodal features based on high-level description of video.
We present a novel motion trajectory based video retrieval system using parallel adaptive self or... more We present a novel motion trajectory based video retrieval system using parallel adaptive self organizing maps (PASOMs) in this paper. The motion trajectories are extracted from video sequences by a robust tracker using detection-based particle filter. To reduce the high dimension of motion trajectories and achieve flexible video retrieval, our algorithm first decomposes each raw motion trajectory into subtrajectories by using a maximum acceleration based approach. Each subtrajectory is then modeled and coded by two different methods, polynomial curving fitting and independent component analysis. To fuse the different features of subtrajectories for more efficient and flexible retrieval, we use a PASOMs network as the searching tool where each SOM learns and classifies one feature. Experimental results show the superior performance of the proposed approach for video retrieval comparing with prior approaches.
2003-04-12). http://multimedia. eecs. uic. edu/fai- …, Jan 1, 2003
This paper explores the domain of automatically extracting semantic information from given databa... more This paper explores the domain of automatically extracting semantic information from given database schema and structured textual data for the purpose of Ontology construction. We examine the advantages of ontology based information search mechanisms and survey the difference techniques currently employed for automated ontology construction from given text corpus and from unstructured multimedia data, such as images, audio and video. Finally we propose an algorithm for ontology construction from given structured textual data with the help of database schema. The approach that we thoroughly discuss is from the domain of Formal Concept Analysis which is based on Lattice Theory. We use a generic lexical ontology WordNet to generate word senses and to label the nodes of ontology.
… , 2004. Proceedings of INMIC 2004. 8th …, Jan 1, 2004
In this paper, we outline a novel technique fbr robust indexing and refrieval of video object frn... more In this paper, we outline a novel technique fbr robust indexing and refrieval of video object frnjectoi-ies. Our approach is based on using sputiaternporul crrrilature to represent the trajectories. Daminai7t inflection poirits ,from the czrrvotrr?'e ure ex/racred at mirltiple levels of scale simultaneously reveulii.rg sinrctrtre ai vnwing levels o j details. These inyeclion poiiils, the maxima of Curvattire Scale Space (CSS}, are then used Jbr indexing urd retrieval. E.rpeuirnents 017 a dalahase containing trajectories of high jumps crnd other motions show the ufiirieinvarionl nolure of o w approach.
This figure compares our system in terms of stand-off distance and capture volume with others in ... more This figure compares our system in terms of stand-off distance and capture volume with others in recent literature.
This paper presents novel tensor-based object trajectory modelling techniques for simultaneous re... more This paper presents novel tensor-based object trajectory modelling techniques for simultaneous representation of multiple objects motion trajectories in a content based indexing and retrieval framework. Three different tensor decomposition techniques-PARAFAC, HOSVD and Multiple-SVD-are explored to achieve this goal with the aim of using a minimum set of coefficients and data-dependant bases. These tensor decompositions have been applied to represent full as well as segmented trajectories. Our simulation results show that the PARAFAC-based representation provides higher compression ratio, superior precision-recall metrics, and smaller query processing time compared to the other tensor-based approaches.
Multimedia data, such as text, audio, images and video, is rapidly evolving as the main form for ... more Multimedia data, such as text, audio, images and video, is rapidly evolving as the main form for the creation, exchange, and storage of information in the modern era. Primarily, this is attributed to rapid advances in the three major technologies that determine its growth: VLSI technology that is producing greater processing power; broad-band networks (ISDN, ATM, etc) that are providing much higher bandwidth for many practical applications, and multimedia compression standards (JPEG, H.263, MPEG, MP3, etc) that enable efficient storage and communication. The combination of these three advances is spurring the creation and processing of increasingly high-volume multimedia data, along with its efficient compression and transmission over high-bandwidth networks. This current trend towards the removal of any conceivable bottleneck in using multimedia and its impact on a whole spectrum of users, from advanced research organizations to home users, has led to the explosive growth of visual information available in the form of digital libraries and online multimedia archives. According to a press release by Google Inc. in December 2001, the search engine offers access to over 3 billion web documents and its Image search comprises more that 330 million images. AltaVista has been serving around 25 million search queries per day in more than 25 languages, with its multimedia search featuring over 45 million images, videos and audio clips.
International Workshop on Frontiers of …, Jan 1, 2003
This paper provides an overview of different video content modeling techniques employed in existi... more This paper provides an overview of different video content modeling techniques employed in existing content-based video indexing and retrieval (CBVIR) systems. Based on the modeling requirements of a hypothetical (somewhat ideal) CBVIR system, we analyze and categorize existing modeling approaches. Starting with a review of techniques to model raw video data, we study approaches used to describe physical objects, and conclude with a review on high-level semantic modeling of data with focus on the multimodal analysis. Based on the current status of research in CBVIR systems, we identify the growth potential, future directions, and open research issues. Finally, a hypothetical CBVIR system is outlined in the concluding remarks, which exploits object-based representation of MPEG-4 compressed bitstream and uses multimodal features based on high-level description of video.
We present a compressed domain video object segmentation method for the MPEG encoded video sequen... more We present a compressed domain video object segmentation method for the MPEG encoded video sequences. For a fraction of the raw domain analysis, compressed domain segmentation provides the essential priori information to many vision tasks from surveillance to transcoding that require fast processing of large volumes of data where pixel-resolution boundary extraction is not required. Our method generates accurate segmentation maps in block-resolution at hierarchically varying object levels, which empowers application to determine the most pertinent partition of images. It exploits the block-structure of the compressed video to minimize the amount of data to be processed. All the available motion flow within a group-of-picture are projected onto a single layer that also consists of the frequency decomposition of color pattern. Then, by starting from the blocks where the spatial energy is small, it expands homogeneous regions while automatically adapting local similarity criteria. We also formulate an alternative solution that applies a kernel based clustering where separate spatial, transform, and motion kernels are used to establish the affinity. We show that both region expansion and mean-shift produce similar results as the computationally expensive raw domain segmentation. Finally, a binary clustering iteratively merges the most similar regions to generate a hierarchical partition tree.
Visual Communications and Image Processing, Jan 1, 2007
This paper addresses the issue of multi-source collaborative object tracking in high-definition (... more This paper addresses the issue of multi-source collaborative object tracking in high-definition (HD) video sequences. Specifically, we propose a new joint tracking paradigm for the multiple stream electronic pan-tilt-zoom (EPTZ) cameras. These cameras are capable of transmitting a low resolution thumbnail (LRT) image of the whole field of view as well as a high-resolution cropped (HRC) image for the target region. We exploit this functionality to perform joint tracking in both low resolution image of the whole field of view as well as high resolution image of the moving target. Our system detects objects of interest in the LRT image by background subtraction and tracks them using iterative coupled refinement in both LRT and HRC images. We compared the performance of our joint tracking system with that of tracking only in the HD mode. The results of our experiments show improved performance in terms of higher frame rates and better localization.
Motion information is regarded as one of the most important cues for developing semantics in vide... more Motion information is regarded as one of the most important cues for developing semantics in video data. Yet it is extremely challenging to build semantics of video clips particularly when it involves interactive motion of multiple objects. Most of the existing research has focused on capturing and modelling the motion of each object individually thus loosing interaction information. Such approaches yield low precisionrecall ratios and limited indexing and retrieval performances. This paper presents a novel framework for compact representation of multi-object motion trajectories. Three efficient multitrajectory indexing and retrieval algorithms based on multilinear algebraic representations are proposed. These include: (i) geometrical multiple-trajectory indexing and retrieval (GMIR), (ii) unfolded multiple-trajectory indexing and retrieval (UMIR), and (iii) concentrated multiple-trajectory indexing and retrieval (CMIR). The proposed tensor-based representations not only remarkably reduce the dimensionality of the indexing space but also enable the realization of fast retrieval systems. The proposed algorithms have been implemented and evaluated using real video datasets. Simulation results on both full and partial (segmented) multiple motion trajectories with varying number of objects, trajectory lengths, and sampling rates demonstrate that the CMIR algorithm provides superior precision-recall metrics, and smaller query processing time compared to the other approaches.
In recent years the human iris has established itself as a robust, highly accurate biometric. Aca... more In recent years the human iris has established itself as a robust, highly accurate biometric. Academic research of iris biometrics is expanding and several iris recognition products have become commercially available. Conventional iris recognition systems impose constraints with regard to a ...
In this paper, we propose a novel technique for modelbased recognition of complex object motion t... more In this paper, we propose a novel technique for modelbased recognition of complex object motion trajectories using Gaussian Mixture Models (GMM). We build our models on Principal Component Analysis (PCA)-based representation of trajectories after segmenting them into small units of perceptually similar pieces of motions. These subtrajectories are then fitted with automaticallylearnt mixture of Gaussians to estimate the underlying class probability distribution. Experiments are performed on two data sets; the ASL data set (from UCI's KDD archives) consists of 207 trajectories depicting signs for three words, from Australian Sign Language (ASL); the HJSL data set contains 108 trajectories from sports videos. Our experiments yield an accuracy of 85+% performing much better than existing approaches.
Motion trajectories provide rich spatio-temporal information about an object's activity. The traj... more Motion trajectories provide rich spatio-temporal information about an object's activity. The trajectory information can be obtained using a tracking algorithm on data streams available from a range of devices including motion sensors, video cameras, haptic devices, etc. Developing view-invariant activity recognition algorithms based on this high dimensionality cue is an extremely challenging task. This paper presents efficient activity recognition algorithms using novel view-invariant representation of trajectories thus will recognize motion activities independent of the view point. Towards this end, we derive two Affine-invariant representations for motion trajectories based on curvature scale space (CSS) and centroid distance function (CDF). The properties of these schemes facilitate the design of efficient recognition algorithms based on hidden Markov models (HMMs). In the CSS-based representation, maxima of curvature zero crossings at increasing levels of smoothness are extracted to mark the location and extent of concavities in the curvature. The sequences of these CSS maxima are then modeled by continuous density (HMMs). For the case of CDF, we first segment the trajectory into subtrajectories using CDF-based representation. These subtrajectories are then represented by their Principal Component Analysis (PCA) coefficients. The sequences of these PCA coefficients from subtrajectories are then modeled by continuous density hidden Markov models (HMMs). Different classes of object motions are modeled by one Continuous HMM per class where state PDFs are represented by GMMs. Experiments using a database of around 1750 complex trajectories (obtained from UCI-KDD data archives) subdivided into 5 different classes are reported.
Abstract This paper studies efficient feature spaces for content based indexing and retrieval of ... more Abstract This paper studies efficient feature spaces for content based indexing and retrieval of object motion trajectories. Taking object trajectory data as input, we first investigate highly compact affine invariant feature spaces based on Fourier Descriptors (FD) and Principal Component Analysis (PCA) techniques. Based on these feature spaces, we then develop a hybrid content based indexing and retrieval system that employs a two-stage matching scheme. The first stage uses affine-invariant Fourier Descriptor (FD) for indexing and ...
In this paper, we propose a novel technique for modelbased recognition of complex object motion t... more In this paper, we propose a novel technique for modelbased recognition of complex object motion trajectories using Hidden Markov Models (HMM). We build our models on Principal Component Analysis (PCA)-based representation of trajectories after segmenting them into small units of perceptually similar pieces of motions. These subtrajectories are then grouped using spectral clustering to decide on the number of states for each HMM representing a class of object motion. The hidden states of the HMMs are represented by Gaussian Mixtures (GM's). This way the HMM topology as well as the parameters are automatically derived from the training data in a fully unsupervised sense. Experiments are performed on two data sets; the ASL data set (from UCI's KDD archives) consists of 207 trajectories depicting signs for three words, from Australian Sign Language (ASL); the HJSL data set contains 108 trajectories from sports videos. Our experiments yield an accuracy of 90+% performing much better than existing approaches.
Uploads
Papers by Faisal Bashir