This paper considers the framework in which World-Wide-Web based online courseware can be develop... more This paper considers the framework in which World-Wide-Web based online courseware can be developed. This framework is illustrated using courseware developed for parallel computing, C programming, X-Window programming and computer vision.
Boundary representation models reverse engineered from 3D range data suffer from various inaccura... more Boundary representation models reverse engineered from 3D range data suffer from various inaccuracies caused by noise in the measured data and the model building software. Beautification aims to improve such models in a post-processing step solely working with the boundary representation model. The improved model should exhibit topological and geometric regularities representing the original, ideal design intent. This paper gives an overview of algorithms for a complete beautification system suitable for improving the topology and the geometry of low to medium complexity reverse engineered models.
Proceedings of the sixth ACM symposium on Solid modeling and applications, 2001
The authors are developing an automated reverse engineering system for reconstructing the shape o... more The authors are developing an automated reverse engineering system for reconstructing the shape of simple mechanical parts. B-rep models are created by fitting surfaces to point clouds obtained by scanning an object using a 3D laser scanner. The resulting models, although valid, are often not suitable for purposes such as redesign because expected regularities and constraints are not present. This information is lost because each face of the model is determined independently. A global approach is required, in particular one that is capable of finding symmetries originally present. This paper describes a practical algorithm for finding global symmetries in suitable B-rep models built from planes, spheres, cylinders, cones and tori. It has been implemented and used to determine approximate symmetries of models with up to about 200 vertices in reasonable time. The time performance of the algorithm in the worst case is bounded by O(n 3.5 log 4 n), and a justification is given that on common engineering objects it takes about O(n 2 log 4 n), making it a practical tool for use in a reverse engineering package. Details of the algorithm are given, along with some results from a number of illustrative test runs.
Bas-reliefs are a form of flattened artwork, part-way between 3D sculpture and 2D painting. Recen... more Bas-reliefs are a form of flattened artwork, part-way between 3D sculpture and 2D painting. Recent research has considered automatic bas-relief generation from 3D scenes. However, little work has addressed the generation of bas-reliefs from 2D images. In this paper, we propose a method to automatically generate bas-relief surfaces from frontal photographs of human faces, with potential applications to e.g. coinage and commemorative medals. Our method has two steps. Starting from a photograph of a human face, we first generate a plausible image of a bas-relief of the same face. Secondly, we apply shape-from-shading to this generated bas-relief image to determine the 3D shape of the final bas-relief. To model the mapping from an input photograph to the image of a corresponding bas-relief, we use a feedforward network. The training data comprises images generated from an input 3D model of a face, and images generated from a corresponding bas-relief; the latter is produced by an existing 3D model-to-bas-relief algorithm. A saliency map of the face controls both model building, and bas-relief generation. Our experimental results demonstrate that the generated bas-relief surfaces are smooth and plausible, with correct global geometric nature, the latter giving them a stable appearance under changes of viewing direction and illumination.
Objective-To establish three-dimensional (3D) reference data on average lip movement in normal he... more Objective-To establish three-dimensional (3D) reference data on average lip movement in normal healthy subjects using statistical shape analysis techniques.
Orthognathic surgery is a well-documented clinical procedure for patients who complain of functio... more Orthognathic surgery is a well-documented clinical procedure for patients who complain of functional and aesthetic problems related to dento-facial imbalance. The aims of treatment are to establish facial harmony, normalise function and maximise the stability of the outcome. Conventional diagnostic aids to treatment include plain film radiographs and photographs which provide a static analysis of the facial structures at a particular point in time. Functional assessment of facial movement is rarely performed but could have important implications on the stability of the surgery through post-surgical soft tissue adaptation to the new skeletal relationships. A case report is presented for which novel three-dimensional motion analysis techniques were used to assess facial movement through the surgical correction of a Class 3 malocclusion. It was found that facial movement increased for all circum-oral postures post-surgery. Future research in this field recommends collection of data from a control group to allow comparisons of facial movement between different surgical samples.
Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004., 2004
We present a hierarchical image based facial model which is driven from speech. It incorporates a... more We present a hierarchical image based facial model which is driven from speech. It incorporates a novel modelling and synthesis algorithm for learning and producing coarticulated mouth animation We demonstrate using the hierarchical model how animation of the entire face may be created solely from animations of the mouth and how colour may be reincorporated and reproduced compactly without explicitly being modelled. We then consider and evaluate methods of merging animations from several different facial areas for delivery of output.
IEEE Transactions on Visualization and Computer Graphics, 2013
We introduce a video based approach for producing water surface models. Recent advances in this f... more We introduce a video based approach for producing water surface models. Recent advances in this field output high quality results but require dedicated capturing devices and only work in limited conditions. In contrast, our method achieves a good trade off between the visual quality and the production cost: it automatically produces a visually plausible animation using a single viewpoint video as the input. Our approach is based on two discoveries: first, shape from shading (SFS) is adequate to capture the appearance and dynamic behaviour of the example water; second, shallow water model can be used to estimate a velocity field that produces complex surface dynamics. We will provide qualitative evaluation of our method and demonstrate its good performance across a wide range of scenes.
IEEE Transactions on Visualization and Computer Graphics, 2013
3D surface registration transforms multiple 3D datasets into the same coordinate system so as to ... more 3D surface registration transforms multiple 3D datasets into the same coordinate system so as to align overlapping components of these sets. Recent surveys have covered different aspects of either rigid or non-rigid registration, but seldom discuss them as a whole. Our study serves two purposes: (i) to give a comprehensive survey of both types of registration, focusing on 3D point clouds and meshes, and (ii) to provide a better understanding of registration from the perspective of data fitting. Registration is closely related to data fitting in that it comprises three core interwoven components: model selection, correspondences & constraints and optimization. Study of these components (i) provides a basis for comparison of the novelties of different techniques, (ii) reveals the similarity of rigid and non-rigid registration in terms of problem representations, and (iii) shows how over-fitting arises in non-rigid registration and the reasons for increasing interest in intrinsic techniques. We further summarise some practical issues of registration which include initializations and evaluations, and discuss some of our own observations, insights and foreseeable research trends.
Procedings of the British Machine Vision Conference 1999, 1999
This paper provides two algorithms; one for adding eigenspaces, another for subtracting them, thu... more This paper provides two algorithms; one for adding eigenspaces, another for subtracting them, thus allowing for incremental updating and downdating of data models. Importantly, and unlike previous work, we keep an accurate track of the mean of the data, which allows our methods to be used in classification applications. The result of adding eigenspaces, each made from a set of data, is an approximation to that which would obtain were the sets of data taken together. Subtracting eigenspaces yields a result approximating that which would obtain were a subset of data used. Using our algorithms it is possible to perform "arithmetic" on eigenspaces without reference to the original data. We illustrate the use of our algorithms in three generic applications, including the dynamic construction of Gaussian mixture models.
This paper provides algorithms for adding and subtracting eigenspaces, thus allowing for incremen... more This paper provides algorithms for adding and subtracting eigenspaces, thus allowing for incremental updating and downdating of data models. Importantly, and unlike previous work, we keep an accurate track of the mean of the data, which allows our methods to be used in classification applications. The result of adding eigenspaces, each made from a set of data, is an approximation to that which would obtain were the sets of data taken together. Subtracting eigenspaces yields a result approximating that which would obtain were a subset of data used. Using our algorithms, it is possible to perform 'arithmetic' on eigenspaces without reference to the original data. Eigenspaces can be constructed using either eigenvalue decomposition (EVD) or singular value decomposition (SVD). We provide addition operators for both methods, but subtraction for EVD only, arguing there is no closed-form solution for SVD. The methods and discussion surrounding SVD provide the principle novelty in this paper. We illustrate the use of our algorithms in three generic applications, including the dynamic construction of Gaussian mixture models.
In this paper we describe a method for re-mapping animation parameters between multiple types of ... more In this paper we describe a method for re-mapping animation parameters between multiple types of facial model for performance driven animation. A facial performance can be analysed automatically in terms of a set of facial action trajectories using a modified appearance model with modes of variation encoding specific facial actions. These parameters can then be used to animate other appearance models, or 3D facial models. Thus, the animation parameters analysed from the video performance may be re-used to animate multiple types of facial model. We demonstrate the effectiveness of our approach by measuring its ability to successfully extract action-parameters from performances and by displaying frames from example animations.
Automatic facial expression analysis aims to analyse human facial expressions and classify them i... more Automatic facial expression analysis aims to analyse human facial expressions and classify them into discrete categories. Methods based on existing work are reliant on extracting information from video sequences and employ either some form of subjective thresholding of dynamic information or attempt to identify the particular individual frames in which the expected behaviour occurs. These methods are inefficient as they require either additional subjective information, tedious manual work or fail to take advantage of the information contained in the dynamic signature from facial movements for the task of expression recognition. In this paper, a novel framework is proposed for automatic facial expression analysis which extracts salient information from video sequences but does not rely on any subjective preprocessing or additional user-supplied information to select frames with peak expressions. The experimental framework demonstrates that the proposed method outperforms static expression recognition systems in terms of recognition rate. The approach does not rely on action units (AUs) and therefore, eliminates errors which are otherwise propagated to the final result due to incorrect initial identification of AUs. The proposed framework explores a parametric space of over 300 dimensions and is tested with six state-of-the-art machine learning techniques. Such robust and extensive experimentation provides an important foundation for the assessment of the performance for future work. A further contribution of the paper is offered in the form of a user study. This was conducted in order to investigate the correlation between human cognitive systems and the proposed framework for the understanding of human emotion classification and the reliability of public databases.
Procedings of the British Machine Vision Conference 2003, 2003
In this paper we present a method for adding Hidden Markov Models. The main advantages of our met... more In this paper we present a method for adding Hidden Markov Models. The main advantages of our method are that it does not require the data the models had been trained on, allows a change in the number of components, does not assume independence of the components to be added and is resistant to the order in which the training data arrives. We assessed the method in the experiments with synthetic data, which showed good accuracy. Finally, we present an application in computer vision.
2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2012
We consider the problem of producing accurate facial animation corresponding to a given input spe... more We consider the problem of producing accurate facial animation corresponding to a given input speech signal. A popular technique previously used for Audio Driven Facial Animation is to build a joint audiovisual model using Active Appearance Models (AAMs) to represent possible facial variations and Hidden Markov Models (HMMs) to select the correct appearance based on the input audio. However there are several questions that remained unanswered. In particular the choice of clustering technique and the choice of the number of clusters in the HMM may have significant influence over the quality of the produced videos. We have investigated a range of clustering techniques in order to improve the quality of the HMM produced, and proposed a new structure based on using Gaussian Mixture Models (GMMs) to model each phoneme separately. We compared our approach to several alternatives using a public dataset of 300 phonetically labeled sentences spoken by a single person and found that our approach produces more accurate animation. In addition, we use a hybrid approach where the training data is phonetically labeled thus producing a model with better separation of phonemes, but test audio data is not labeled, thus making our approach for generating facial animation less laborious and fully automatic.
This paper describes a novel approach to building models of temporal dynamics for facial animatio... more This paper describes a novel approach to building models of temporal dynamics for facial animation with applications in performing perceptual testing of trustworthiness. A vital component of the system is a method to bring two image sequences into temporal alignment. Our approach is to project the two sequences into face space (built using shape models [1]) and apply dynamic time warping (DTW). However, the variability in the sequences causes the standard DTW algorithm to perform poorly on our data, and so we have overcome this by extending DTW in the following ways: 1) the signal magnitudes are augmented by incorporating derivatives [2], and a scheme for estimating weights in the cost function is proposed, 2) the set of sequences is used to build a graph, with nodes representing sequences and edges indicating the cost of applying the extended DTW to align pairs of sequences; better alignments between sequences can now be found by traversing the minimum cost path through the graph. Once all signals are aligned to a common temporal reference it is straightforward to map the temporal dynamics from one face to another. A remapped face is synthesised using the new trajectory in face space to drive an active appearance model [1]. Furthermore, the common temporal reference allows us to build a statistical model of the dynamics. This can be used to both identify dynamics of interest and also to manipulate the dynamics, e.g. to reduce or exaggerate facial dynamics.
In this paper, we address the problem of automating the partial representation from real world da... more In this paper, we address the problem of automating the partial representation from real world data with an unknown a priori structure. Such representation could be very useful for the further construction of an automatic hierarchical data model. We propose a three stage process using data normalisation and the data intrinsic dimensionality estimation as the first step. The second stage uses a modified sparse Non-negative matrix factorization (sparse NMF) algorithm to perform the initial segmentation. At the final stage region growing algorithm is applied to construct a mask of the original data. Our algorithm has a very broad range of a potential applications, we illustrate this versatility by applying the algorithm to several dissimilar data sets.
3rd European Conference on Visual Media Production (CVMP 2006). Part of the 2nd Multimedia Conference 2006, 2006
ABSTRACT In this paper, we present a novel approach for generating a variety of complex behaviour... more ABSTRACT In this paper, we present a novel approach for generating a variety of complex behaviours. Our approach is model-based. We train a dual hidden Markov model (HMM) on 3D motion capture (MoCap) data representing a number of interactions between two people. Then we track 3D motion of a person in ordinary 2D video. Finally, using the dual HMM and the Viterbi algorithm, we generate a moving "virtual friend" reacting to the given motion of tracked person.
This paper considers the framework in which World-Wide-Web based online courseware can be develop... more This paper considers the framework in which World-Wide-Web based online courseware can be developed. This framework is illustrated using courseware developed for parallel computing, C programming, X-Window programming and computer vision.
Boundary representation models reverse engineered from 3D range data suffer from various inaccura... more Boundary representation models reverse engineered from 3D range data suffer from various inaccuracies caused by noise in the measured data and the model building software. Beautification aims to improve such models in a post-processing step solely working with the boundary representation model. The improved model should exhibit topological and geometric regularities representing the original, ideal design intent. This paper gives an overview of algorithms for a complete beautification system suitable for improving the topology and the geometry of low to medium complexity reverse engineered models.
Proceedings of the sixth ACM symposium on Solid modeling and applications, 2001
The authors are developing an automated reverse engineering system for reconstructing the shape o... more The authors are developing an automated reverse engineering system for reconstructing the shape of simple mechanical parts. B-rep models are created by fitting surfaces to point clouds obtained by scanning an object using a 3D laser scanner. The resulting models, although valid, are often not suitable for purposes such as redesign because expected regularities and constraints are not present. This information is lost because each face of the model is determined independently. A global approach is required, in particular one that is capable of finding symmetries originally present. This paper describes a practical algorithm for finding global symmetries in suitable B-rep models built from planes, spheres, cylinders, cones and tori. It has been implemented and used to determine approximate symmetries of models with up to about 200 vertices in reasonable time. The time performance of the algorithm in the worst case is bounded by O(n 3.5 log 4 n), and a justification is given that on common engineering objects it takes about O(n 2 log 4 n), making it a practical tool for use in a reverse engineering package. Details of the algorithm are given, along with some results from a number of illustrative test runs.
Bas-reliefs are a form of flattened artwork, part-way between 3D sculpture and 2D painting. Recen... more Bas-reliefs are a form of flattened artwork, part-way between 3D sculpture and 2D painting. Recent research has considered automatic bas-relief generation from 3D scenes. However, little work has addressed the generation of bas-reliefs from 2D images. In this paper, we propose a method to automatically generate bas-relief surfaces from frontal photographs of human faces, with potential applications to e.g. coinage and commemorative medals. Our method has two steps. Starting from a photograph of a human face, we first generate a plausible image of a bas-relief of the same face. Secondly, we apply shape-from-shading to this generated bas-relief image to determine the 3D shape of the final bas-relief. To model the mapping from an input photograph to the image of a corresponding bas-relief, we use a feedforward network. The training data comprises images generated from an input 3D model of a face, and images generated from a corresponding bas-relief; the latter is produced by an existing 3D model-to-bas-relief algorithm. A saliency map of the face controls both model building, and bas-relief generation. Our experimental results demonstrate that the generated bas-relief surfaces are smooth and plausible, with correct global geometric nature, the latter giving them a stable appearance under changes of viewing direction and illumination.
Objective-To establish three-dimensional (3D) reference data on average lip movement in normal he... more Objective-To establish three-dimensional (3D) reference data on average lip movement in normal healthy subjects using statistical shape analysis techniques.
Orthognathic surgery is a well-documented clinical procedure for patients who complain of functio... more Orthognathic surgery is a well-documented clinical procedure for patients who complain of functional and aesthetic problems related to dento-facial imbalance. The aims of treatment are to establish facial harmony, normalise function and maximise the stability of the outcome. Conventional diagnostic aids to treatment include plain film radiographs and photographs which provide a static analysis of the facial structures at a particular point in time. Functional assessment of facial movement is rarely performed but could have important implications on the stability of the surgery through post-surgical soft tissue adaptation to the new skeletal relationships. A case report is presented for which novel three-dimensional motion analysis techniques were used to assess facial movement through the surgical correction of a Class 3 malocclusion. It was found that facial movement increased for all circum-oral postures post-surgery. Future research in this field recommends collection of data from a control group to allow comparisons of facial movement between different surgical samples.
Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004., 2004
We present a hierarchical image based facial model which is driven from speech. It incorporates a... more We present a hierarchical image based facial model which is driven from speech. It incorporates a novel modelling and synthesis algorithm for learning and producing coarticulated mouth animation We demonstrate using the hierarchical model how animation of the entire face may be created solely from animations of the mouth and how colour may be reincorporated and reproduced compactly without explicitly being modelled. We then consider and evaluate methods of merging animations from several different facial areas for delivery of output.
IEEE Transactions on Visualization and Computer Graphics, 2013
We introduce a video based approach for producing water surface models. Recent advances in this f... more We introduce a video based approach for producing water surface models. Recent advances in this field output high quality results but require dedicated capturing devices and only work in limited conditions. In contrast, our method achieves a good trade off between the visual quality and the production cost: it automatically produces a visually plausible animation using a single viewpoint video as the input. Our approach is based on two discoveries: first, shape from shading (SFS) is adequate to capture the appearance and dynamic behaviour of the example water; second, shallow water model can be used to estimate a velocity field that produces complex surface dynamics. We will provide qualitative evaluation of our method and demonstrate its good performance across a wide range of scenes.
IEEE Transactions on Visualization and Computer Graphics, 2013
3D surface registration transforms multiple 3D datasets into the same coordinate system so as to ... more 3D surface registration transforms multiple 3D datasets into the same coordinate system so as to align overlapping components of these sets. Recent surveys have covered different aspects of either rigid or non-rigid registration, but seldom discuss them as a whole. Our study serves two purposes: (i) to give a comprehensive survey of both types of registration, focusing on 3D point clouds and meshes, and (ii) to provide a better understanding of registration from the perspective of data fitting. Registration is closely related to data fitting in that it comprises three core interwoven components: model selection, correspondences & constraints and optimization. Study of these components (i) provides a basis for comparison of the novelties of different techniques, (ii) reveals the similarity of rigid and non-rigid registration in terms of problem representations, and (iii) shows how over-fitting arises in non-rigid registration and the reasons for increasing interest in intrinsic techniques. We further summarise some practical issues of registration which include initializations and evaluations, and discuss some of our own observations, insights and foreseeable research trends.
Procedings of the British Machine Vision Conference 1999, 1999
This paper provides two algorithms; one for adding eigenspaces, another for subtracting them, thu... more This paper provides two algorithms; one for adding eigenspaces, another for subtracting them, thus allowing for incremental updating and downdating of data models. Importantly, and unlike previous work, we keep an accurate track of the mean of the data, which allows our methods to be used in classification applications. The result of adding eigenspaces, each made from a set of data, is an approximation to that which would obtain were the sets of data taken together. Subtracting eigenspaces yields a result approximating that which would obtain were a subset of data used. Using our algorithms it is possible to perform "arithmetic" on eigenspaces without reference to the original data. We illustrate the use of our algorithms in three generic applications, including the dynamic construction of Gaussian mixture models.
This paper provides algorithms for adding and subtracting eigenspaces, thus allowing for incremen... more This paper provides algorithms for adding and subtracting eigenspaces, thus allowing for incremental updating and downdating of data models. Importantly, and unlike previous work, we keep an accurate track of the mean of the data, which allows our methods to be used in classification applications. The result of adding eigenspaces, each made from a set of data, is an approximation to that which would obtain were the sets of data taken together. Subtracting eigenspaces yields a result approximating that which would obtain were a subset of data used. Using our algorithms, it is possible to perform 'arithmetic' on eigenspaces without reference to the original data. Eigenspaces can be constructed using either eigenvalue decomposition (EVD) or singular value decomposition (SVD). We provide addition operators for both methods, but subtraction for EVD only, arguing there is no closed-form solution for SVD. The methods and discussion surrounding SVD provide the principle novelty in this paper. We illustrate the use of our algorithms in three generic applications, including the dynamic construction of Gaussian mixture models.
In this paper we describe a method for re-mapping animation parameters between multiple types of ... more In this paper we describe a method for re-mapping animation parameters between multiple types of facial model for performance driven animation. A facial performance can be analysed automatically in terms of a set of facial action trajectories using a modified appearance model with modes of variation encoding specific facial actions. These parameters can then be used to animate other appearance models, or 3D facial models. Thus, the animation parameters analysed from the video performance may be re-used to animate multiple types of facial model. We demonstrate the effectiveness of our approach by measuring its ability to successfully extract action-parameters from performances and by displaying frames from example animations.
Automatic facial expression analysis aims to analyse human facial expressions and classify them i... more Automatic facial expression analysis aims to analyse human facial expressions and classify them into discrete categories. Methods based on existing work are reliant on extracting information from video sequences and employ either some form of subjective thresholding of dynamic information or attempt to identify the particular individual frames in which the expected behaviour occurs. These methods are inefficient as they require either additional subjective information, tedious manual work or fail to take advantage of the information contained in the dynamic signature from facial movements for the task of expression recognition. In this paper, a novel framework is proposed for automatic facial expression analysis which extracts salient information from video sequences but does not rely on any subjective preprocessing or additional user-supplied information to select frames with peak expressions. The experimental framework demonstrates that the proposed method outperforms static expression recognition systems in terms of recognition rate. The approach does not rely on action units (AUs) and therefore, eliminates errors which are otherwise propagated to the final result due to incorrect initial identification of AUs. The proposed framework explores a parametric space of over 300 dimensions and is tested with six state-of-the-art machine learning techniques. Such robust and extensive experimentation provides an important foundation for the assessment of the performance for future work. A further contribution of the paper is offered in the form of a user study. This was conducted in order to investigate the correlation between human cognitive systems and the proposed framework for the understanding of human emotion classification and the reliability of public databases.
Procedings of the British Machine Vision Conference 2003, 2003
In this paper we present a method for adding Hidden Markov Models. The main advantages of our met... more In this paper we present a method for adding Hidden Markov Models. The main advantages of our method are that it does not require the data the models had been trained on, allows a change in the number of components, does not assume independence of the components to be added and is resistant to the order in which the training data arrives. We assessed the method in the experiments with synthetic data, which showed good accuracy. Finally, we present an application in computer vision.
2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2012
We consider the problem of producing accurate facial animation corresponding to a given input spe... more We consider the problem of producing accurate facial animation corresponding to a given input speech signal. A popular technique previously used for Audio Driven Facial Animation is to build a joint audiovisual model using Active Appearance Models (AAMs) to represent possible facial variations and Hidden Markov Models (HMMs) to select the correct appearance based on the input audio. However there are several questions that remained unanswered. In particular the choice of clustering technique and the choice of the number of clusters in the HMM may have significant influence over the quality of the produced videos. We have investigated a range of clustering techniques in order to improve the quality of the HMM produced, and proposed a new structure based on using Gaussian Mixture Models (GMMs) to model each phoneme separately. We compared our approach to several alternatives using a public dataset of 300 phonetically labeled sentences spoken by a single person and found that our approach produces more accurate animation. In addition, we use a hybrid approach where the training data is phonetically labeled thus producing a model with better separation of phonemes, but test audio data is not labeled, thus making our approach for generating facial animation less laborious and fully automatic.
This paper describes a novel approach to building models of temporal dynamics for facial animatio... more This paper describes a novel approach to building models of temporal dynamics for facial animation with applications in performing perceptual testing of trustworthiness. A vital component of the system is a method to bring two image sequences into temporal alignment. Our approach is to project the two sequences into face space (built using shape models [1]) and apply dynamic time warping (DTW). However, the variability in the sequences causes the standard DTW algorithm to perform poorly on our data, and so we have overcome this by extending DTW in the following ways: 1) the signal magnitudes are augmented by incorporating derivatives [2], and a scheme for estimating weights in the cost function is proposed, 2) the set of sequences is used to build a graph, with nodes representing sequences and edges indicating the cost of applying the extended DTW to align pairs of sequences; better alignments between sequences can now be found by traversing the minimum cost path through the graph. Once all signals are aligned to a common temporal reference it is straightforward to map the temporal dynamics from one face to another. A remapped face is synthesised using the new trajectory in face space to drive an active appearance model [1]. Furthermore, the common temporal reference allows us to build a statistical model of the dynamics. This can be used to both identify dynamics of interest and also to manipulate the dynamics, e.g. to reduce or exaggerate facial dynamics.
In this paper, we address the problem of automating the partial representation from real world da... more In this paper, we address the problem of automating the partial representation from real world data with an unknown a priori structure. Such representation could be very useful for the further construction of an automatic hierarchical data model. We propose a three stage process using data normalisation and the data intrinsic dimensionality estimation as the first step. The second stage uses a modified sparse Non-negative matrix factorization (sparse NMF) algorithm to perform the initial segmentation. At the final stage region growing algorithm is applied to construct a mask of the original data. Our algorithm has a very broad range of a potential applications, we illustrate this versatility by applying the algorithm to several dissimilar data sets.
3rd European Conference on Visual Media Production (CVMP 2006). Part of the 2nd Multimedia Conference 2006, 2006
ABSTRACT In this paper, we present a novel approach for generating a variety of complex behaviour... more ABSTRACT In this paper, we present a novel approach for generating a variety of complex behaviours. Our approach is model-based. We train a dual hidden Markov model (HMM) on 3D motion capture (MoCap) data representing a number of interactions between two people. Then we track 3D motion of a person in ordinary 2D video. Finally, using the dual HMM and the Viterbi algorithm, we generate a moving "virtual friend" reacting to the given motion of tracked person.
Uploads
Papers by David Marshall