Dr Teo de Campos

Research Fellow

Qualifications: DPhil in Engineering Sciences, Oxford, 2006

Email:
Phone: Work: 01483 68 4705
Room no: 03 BB 00

Further information

Biography

I'm currently focusing my efforts on the ACASVA project, supporting several aspects of this project, in particular working with action recognition and transfer learning, with my PhD student Nazli FarajiDavar. I work closely with Bill Christmas, Josef Kittler and others. In the past, I've also collaborated with Mark Barnard and Krystian Mikolajczyk.

Here is a summary of activities in my previous jobs and courses:

  • At Xerox, I was the main investigator of a highly collaborative work package (WP) of the PinView project. This WP was about image categorisation using visual attention information. Among other things, this work resulted in a paper published at CVIU. I also supervised Julian McAuley's internship and we worked on image region labelling and graph matching. His internship was outstandingly productive: in just 4 months of work, we submitted two patent applications and two papers for conferences.
  • At Microsoft, I worked with Manik Varma on image classification using multiple types of image descriptors and bags-of-visual-words. One of the outcomes of this work was the Chars74K Character Recognition Dataset.
  • At Sharp, I worked on image processing methods to enhance the image quality on novel LCD technologies, such as multi-view and stereoscopic (3D) displays. I also worked as a member of the "New Project" Project (yes, the word project appears twice :-)).
  • My doctorate in Oxford was about tracking articulated objects in 3D, in real-time. I explored methods of both these approaches: model-based (projective geometry) and discriminative (regression-based). My thesis page gives all the details and shows some demos.
  • I've also done a masters at the University of Sao Paulo (in the Creativision group) on the application of feature selection for face recognition. My thesis was awarded the best MSc thesis by the Brazilian Computer Society in a nationwide contest.

Research Interests

  • transfer learning
  • human action recognition
  • image categorisation and retrieval
  • visual attention

Further details can be found on my personal web page.

Publications

The reference in this page have been generated automatically. This link may be more accurate and it also contains PDFs for download.

Highlights

  • FarajiDavar N, deCampos TE, Kittler J. (2014) 'Adaptive Transductive Transfer Machine'. Nottingham : Preceedings of the British Machine Vision Conference (BMVC),
  • Kittler J, Christmas W, de Campos T, Windridge D, Yan F, Illingworth J, Osman M. (2013) 'Domain Anomaly Detection in Machine Perception: A System Architecture and Taxonomy.'. IEEE Trans Pattern Anal Mach Intell,

    Abstract

    We address the problem of anomaly detection in machine perception. The concept of domain anomaly is introduced as distinct from the conventional notion of anomaly used in the literature. We propose a unified framework for anomaly detection which exposes the multifaceted nature of anomalies and suggest effective mechanisms for identifying and distinguishing each facet as instruments for domain anomaly detection. The framework draws on the Bayesian probabilistic reasoning apparatus which clearly defines concepts such as outlier, noise, distribution drift, novelty detection (object, object primitive), rare events, and unexpected events. Based on these concepts we provide a taxonomy of domain anomaly events. One of the mechanisms helping to pinpoint the nature of anomaly is based on detecting incongruence between contextual and noncontextual sensor(y) data interpretation. The proposed methodology has wide applicability. It underpins in a unified way the anomaly detection applications found in the literature. To illustrate some of its distinguishing features, in here the domain anomaly detection methodology is applied to the problem of anomaly detection for a video annotation system.

  • Sanchez J, Perronnin F, de Campos TE. (2012) 'Modeling the Spatial Layout of Images Beyond Spatial Pyramids'. Pattern Recognition Letters,

    Abstract

    Several state-of-the-art image representations consist in averaging local statistics computed from patch-level descriptors. It has been shown by Boureau et al. that such average statistics suffer from two sources of variance. The first one comes from the fact that a finite set of local statistics are averaged. The second one is due to the variation in the proportion of object-dependent information between different images of the same class. For the problem of object classification, these sources of variance affect negatively the accuracy since they increase the overlap between class-conditional probabilities. Our goal is to include information about the spatial layout of images in image signatures based on average statistics. We show that the traditional approach to including the spatial layout – the Spatial Pyramid (SP) – increases the first source of variance while only weakly reducing the second one. We therefore propose two complementary approaches to account for the spatial layout which are compatible with our goal of variance reduction. The first one models the spatial layout in an image-independent manner (as is the case of the SP) while the second one adapts to the image content. A significant benefit of these approaches with respect to the SP is that they do not incur an increase of the image signature dimensionality. We show on PASCAL VOC 2007, 2008 and 2009 the benefits of our approach.

  • deCampos TE, Csurka G, Perronnin F. (2012) 'Images as Sets of Locally Weighted Features'. Computer Vision and Image Understanding (CVIU), 116 (1), pp. 68-85.

    Abstract

    This paper presents a generic framework in which images are modelled as order-less sets of weighted visual features. Each visual feature is associated with a weight factor that may inform its relevance. This framework can be applied to various bag-of-features approaches such as the bag-of-visual-word or the Fisher kernel representations. We suggest that if dense sampling is used, different schemes to weight local features can be evaluated, leading to results that are often better than the combination of multiple sampling schemes, at a much lower computational cost, because the features are extracted only once. This allows our framework to be a test-bed for saliency estimation methods in image categorisation tasks. We explored two main possibilities for the estimation of local feature relevance. The first one is based on the use of saliency maps obtained from human feedback, either by gaze tracking or by mouse clicks. The method is able to profit from such maps, leading to a significant improvement in categorisation performance. The second possibility is based on automatic saliency estimation methods, including Itti & Koch’s method and SIFT’s DoG. We evaluated the proposed framework and saliency estimation methods using an in house dataset and the PASCAL VOC 2008/2007 dataset, showing that some of the saliency estimation methods lead to a significant performance improvement in comparison to the standard unweighted representation.

  • Faraji Davar N, deCampos TE, Kittler J, Yan F. (2011) 'Transductive Transfer Learning for Action Recognition in Tennis Games'. 3rd International Workshop on Video Event Categorization, Tagging and Retrieval for Real-World Applications (VECTaR), in conjunction with ICCV,
  • De Campos T, Barnard M, Mikolajczyk K, Kittler J, Yan F, Christmas W, Windridge D. (2011) 'An evaluation of bags-of-words and spatio-temporal shapes for action recognition'. 2011 IEEE Workshop on Applications of Computer Vision, WACV 2011, , pp. 344-351.

    Abstract

    Bags-of-visual-Words (BoW) and Spatio-Temporal Shapes (STS) are two very popular approaches for action recognition from video. The former (BoW) is an un-structured global representation of videos which is built using a large set of local features. The latter (STS) uses a single feature located on a region of interest (where the actor is) in the video. Despite the popularity of these methods, no comparison between them has been done. Also, given that BoW and STS differ intrinsically in terms of context inclusion and globality/locality of operation, an appropriate evaluation framework has to be designed carefully. This paper compares these two approaches using four different datasets with varied degree of space-time specificity of the actions and varied relevance of the contextual background. We use the same local feature extraction method and the same classifier for both approaches. Further to BoW and STS, we also evaluated novel variations of BoW constrained in time or space. We observe that the STS approach leads to better results in all datasets whose background is of little relevance to action classification. © 2010 IEEE.

  • McAuley JJ, deCampos TE, Caetano TS. (2010) 'Unified graph matching in Euclidean spaces'. San Francisco : IEEE Conference on Computer Vision and Pattern Recognition (CVPR),
  • McAuley JJ, deCampos T, Csurka G, Perronnin F. (2009) 'Hierarchical Image-Region Labeling via Structured Learning'. London : Proceedings of the British Machine Vision Conference,
  • de Campos TE, Babu BR, Varma M. (2009) 'CHARACTER RECOGNITION IN NATURAL IMAGES'. INSTICC-INST SYST TECHNOLOGIES INFORMATION CONTROL & COMMUNICATION VISAPP 2009: PROCEEDINGS OF THE FOURTH INTERNATIONAL CONFERENCE ON COMPUTER VISION THEORY AND APPLICATIONS, VOL 2, Lisbon, PORTUGAL: 4th International Conference on Computer Vision Theory and Applications, pp. 273-280.

Journal articles

  • Windridge D, Kittler J, de Campos T, Yan F, Christmas W. (2015) 'A Novel Markov Logic Rule Induction Strategy for Characterizing Sports Video Footage'. IEEE MULTIMEDIA, 22 (2), pp. 24-35.
  • Veta M, van Diest PJ, Willems SM, Wang H, Madabhushi A, Cruz-Roa A, Gonzalez F, Larsen ABL, Vestergaard JS, Dahl AB, Cireşan DC, Schmidhuber J, Giusti A, Gambardella LM, Tek FB, Walter T, Wang CW, Kondo S, Matuszewski BJ, Precioso F, Snell V, Kittler J, de Campos TE, Khan AM, Rajpoot NM, Arkoumani E, Lacle MM, Viergever MA, Pluim JPW. (2015) 'Assessment of algorithms for mitosis detection in breast cancer histopathology images'. Medical Image Analysis, 20 (1), pp. 237-248.

    Abstract

    © 2014 Elsevier B.V.The proliferative activity of breast tumors, which is routinely estimated by counting of mitotic figures in hematoxylin and eosin stained histology sections, is considered to be one of the most important prognostic markers. However, mitosis counting is laborious, subjective and may suffer from low inter-observer agreement. With the wider acceptance of whole slide images in pathology labs, automatic image analysis has been proposed as a potential solution for these issues. In this paper, the results from the Assessment of Mitosis Detection Algorithms 2013 (AMIDA13) challenge are described. The challenge was based on a data set consisting of 12 training and 11 testing subjects, with more than one thousand annotated mitotic figures by multiple observers. Short descriptions and results from the evaluation of eleven methods are presented. The top performing method has an error rate that is comparable to the inter-observer agreement among pathologists.

  • Veta M, Diest PJV, Willems SM, Wang H, Madabhushi A, Cruz-Roa A, Gonzalez F, Larsen ABL, Vestergaard JS, Dahl AB, Cireşan DC, Schmidhuber J, Giusti A, Gambardella LM, Tek FB, Walter T, Wang C-W, Kondo S, Matuszewski BJ, Precioso F, Snell V, Kittler J, Campos TED, Khan AM, Rajpoot NM, Arkoumani E, Lacle MM, Viergever MA, Pluim JPW. (2014) 'Assessment of algorithms for mitosis detection in breast cancer histopathology images'.

    Abstract

    The proliferative activity of breast tumors, which is routinely estimated by counting of mitotic figures in hematoxylin and eosin stained histology sections, is considered to be one of the most important prognostic markers. However, mitosis counting is laborious, subjective and may suffer from low inter-observer agreement. With the wider acceptance of whole slide images in pathology labs, automatic image analysis has been proposed as a potential solution for these issues. In this paper, the results from the Assessment of Mitosis Detection Algorithms 2013 (AMIDA13) challenge are described. The challenge was based on a data set consisting of 12 training and 11 testing subjects, with more than one thousand annotated mitotic figures by multiple observers. Short descriptions and results from the evaluation of eleven methods are presented. The top performing method has an error rate that is comparable to the inter-observer agreement among pathologists.

  • Kittler J, Christmas W, de Campos T, Windridge D, Yan F, Illingworth J, Osman M. (2013) 'Domain Anomaly Detection in Machine Perception: A System Architecture and Taxonomy.'. IEEE Trans Pattern Anal Mach Intell,

    Abstract

    We address the problem of anomaly detection in machine perception. The concept of domain anomaly is introduced as distinct from the conventional notion of anomaly used in the literature. We propose a unified framework for anomaly detection which exposes the multifaceted nature of anomalies and suggest effective mechanisms for identifying and distinguishing each facet as instruments for domain anomaly detection. The framework draws on the Bayesian probabilistic reasoning apparatus which clearly defines concepts such as outlier, noise, distribution drift, novelty detection (object, object primitive), rare events, and unexpected events. Based on these concepts we provide a taxonomy of domain anomaly events. One of the mechanisms helping to pinpoint the nature of anomaly is based on detecting incongruence between contextual and noncontextual sensor(y) data interpretation. The proposed methodology has wide applicability. It underpins in a unified way the anomaly detection applications found in the literature. To illustrate some of its distinguishing features, in here the domain anomaly detection methodology is applied to the problem of anomaly detection for a video annotation system.

  • Sanchez J, Perronnin F, de Campos TE. (2012) 'Modeling the Spatial Layout of Images Beyond Spatial Pyramids'. Pattern Recognition Letters,

    Abstract

    Several state-of-the-art image representations consist in averaging local statistics computed from patch-level descriptors. It has been shown by Boureau et al. that such average statistics suffer from two sources of variance. The first one comes from the fact that a finite set of local statistics are averaged. The second one is due to the variation in the proportion of object-dependent information between different images of the same class. For the problem of object classification, these sources of variance affect negatively the accuracy since they increase the overlap between class-conditional probabilities. Our goal is to include information about the spatial layout of images in image signatures based on average statistics. We show that the traditional approach to including the spatial layout – the Spatial Pyramid (SP) – increases the first source of variance while only weakly reducing the second one. We therefore propose two complementary approaches to account for the spatial layout which are compatible with our goal of variance reduction. The first one models the spatial layout in an image-independent manner (as is the case of the SP) while the second one adapts to the image content. A significant benefit of these approaches with respect to the SP is that they do not incur an increase of the image signature dimensionality. We show on PASCAL VOC 2007, 2008 and 2009 the benefits of our approach.

  • deCampos TE, Csurka G, Perronnin F. (2012) 'Images as Sets of Locally Weighted Features'. Computer Vision and Image Understanding (CVIU), 116 (1), pp. 68-85.

    Abstract

    This paper presents a generic framework in which images are modelled as order-less sets of weighted visual features. Each visual feature is associated with a weight factor that may inform its relevance. This framework can be applied to various bag-of-features approaches such as the bag-of-visual-word or the Fisher kernel representations. We suggest that if dense sampling is used, different schemes to weight local features can be evaluated, leading to results that are often better than the combination of multiple sampling schemes, at a much lower computational cost, because the features are extracted only once. This allows our framework to be a test-bed for saliency estimation methods in image categorisation tasks. We explored two main possibilities for the estimation of local feature relevance. The first one is based on the use of saliency maps obtained from human feedback, either by gaze tracking or by mouse clicks. The method is able to profit from such maps, leading to a significant improvement in categorisation performance. The second possibility is based on automatic saliency estimation methods, including Itti & Koch’s method and SIFT’s DoG. We evaluated the proposed framework and saliency estimation methods using an in house dataset and the PASCAL VOC 2008/2007 dataset, showing that some of the saliency estimation methods lead to a significant performance improvement in comparison to the standard unweighted representation.

  • De Campos TE, Murray DW. (2006) 'Regression-based hand pose estimation from multiple cameras'. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1, pp. 782-789.

    Abstract

    The RVM-based learning method for whole body pose estimation proposed by Agarwal and Triggs is adapted to hand pose recovery. To help overcome the difficulties presented by the greater degree of self-occlusion and the wider range of poses exhibited in hand imagery, the adaptation proposes a method for combining multiple views. Comparisons of performance using single versus multiple views are reported for both synthesized and real imagery, and the effects of the number of image measurements and the number of training samples on performance are explored.1 © 2006 IEEE.

  • de Campos TE, Tordoff BJ, Murray DW. (2006) 'Recovering articulated pose: a comparison of two pre and postimposed constraint methods.'. IEEE Trans Pattern Anal Mach Intell, United States: 28 (1), pp. 163-168.

    Abstract

    We contrast the performance of two methods of imposing constraints during the tracking of articulated objects, the first method preimposing the kinematic constraints during tracking and, thus, using the minimum degrees of freedom, and the second imposing constraints after tracking and, hence, using the maximum. Despite their very different formulations, the methods recover the same pose change. Further comparisons are drawn in terms of computational speed and algorithmic simplicity and robustness, and it is the last area which is the most telling. The results suggest that using built-in constraints is well-suited to tracking individual articulated objects, whereas applying constraints afterward is most suited to problems involving contact and breakage between articulated (or rigid) objects, where the ability to test tracking performance quickly with constraints turned on or off is desirable.

Conference papers

  • Simon Galvez MF, Menzies RD, Fazi FM, de Campos TE, Hilton A. (2015) 'A Listener Position Adaptive Stereo System for Object-Based Reproduction'. Warsaw: 138th Convention of the Audio Engineering Society (138)

    Abstract

    Stereo reproduction of spatial audio allows the creation of stable acoustic images when the listener is placed in the sweet spot, a small region in the vicinity of the axis of symmetry between both loudspeakers. If the listener moves slightly towards one of the sources, however, the images collapse to the loudspeaker the listener is leaning to. In order to overcome such limitation, a stereo reproduction technique that adapts the sweet spot to the listener position is presented here. This strategy introduces a new approach that maximizes listener immersion by rendering object-based audio, in which several audio objects or sources are placed at virtual locations between the stereo span. By using a video tracking device, the listener is allowed to move freely between the loudspeaker span, while loudspeaker outputs are compensated using conventional panning algorithms so that the position of the different audio objects is kept independent from that of the listener.

  • FarajiDavar N, deCampos TE, Kittler J. (2014) 'Transductive Transfer Machine'. Singapore : Preceedings of the Asian Conference on Computer Vision (ACCV),
  • Nunes de Oliveira M, de Campos TE. (2014) 'Datasets Acquisition and Baseline Experiments for Blind Source Separation'. University of Surrey: Science Without Borders Students Workshop
  • FarajiDavar N, deCampos TE, Kittler J. (2014) 'Adaptive Transductive Transfer Machine'. Nottingham : Preceedings of the British Machine Vision Conference (BMVC),
  • deCampos T. (2014) 'A survey on computer vision tools for action recognition, crowd surveillance and suspect retrieval'. Brasilia : XXXIV Congresso da Sociedade Brasileira de Computacao (CSBC), , pp. 1123-1132.
  • Coppi D, De Campos T, Yan F, Kittler J, Cucchiara R. (2014) 'On detection of novel categories and subcategories of images using incongruence'. ICMR 2014 - Proceedings of the ACM International Conference on Multimedia Retrieval 2014, , pp. 337-344.

    Abstract

    Novelty detection is a crucial task in the development of autonomous vision systems. It aims at detecting if samples do not conform with the learnt models. In this paper, we consider the problem of detecting novelty in object recognition problems in which the set of object classes are grouped to form a semantic hierarchy. We follow the idea that, within a semantic hierarchy, novel samples can be defined as samples whose categorization at a specific level contrasts with the categorization at a more general level. This measure indicates if a sample is novel and, in that case, if it is likely to belong to a novel broad category or to a novel sub-category. We present an evaluation of this approach on two hierarchical subsets of the Caltech256 objects dataset and on the SUN scenes dataset, with different classification schemes. We obtain an improvement over Weinshall et al. and show that it is possible to bypass their normalisation heuristic. We demonstrate that this approach achieves good novelty detection rates as far as the conceptual taxonomy is congruent with the visual hierarchy, but tends to fail if this assumption is not satisfied. Copyright 2014 ACM.

  • Dazzi E, deCampos T, Cesar-Jr RM. (2014) 'Improved object matching using structural relations'. IAPR Joint International Workshops on Statistical Techniques in Pattern Recognition and Structural and Synthactic Pattern Recognition (S+SSPR), , pp. 444-453.
  • Campos TED, Khan A, Yan F, Faraji Davar N, Windridge D, Kittler J, Christmas W. (2013) 'A framework for automatic sports video annotation with anomaly detection and transfer learning'. Proceedings of Machine Learning and Cognitive Science, Palma de Mallorca: 3rd EUCogIII Members Conference
  • Calixto I, de Campos TE, Specia L. (2012) 'Images as Context in Statistical Machine Translation'. Sheffield: 2nd Annual Meeting of the EPSRC Network on Vision and Language
  • Almajai I, Yan F, de Campos T, Khan A, Christmas W, Windridge D, Kittler J. (2012) 'Anomaly Detection and Knowledge Transfer in Automatic Sports Video Annotation'. Springer Proceedings of DIRAC Workshop on Detection and Identification of Rare Audivisual Cues, Barcelona, Spain: DIRAC Workshop on Detection and Identification of Rare Audivisual Cues, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases 384, pp. 109-117.

    Abstract

    A key question in machine perception is how to adaptively build upon existing capabilities so as to permit novel functionalities. Implicit in this are the notions of anomaly detection and learning transfer. A perceptual system must firstly determine at what point the existing learned model ceases to apply, and secondly, what aspects of the existing model can be brought to bear on the newlydefined learning domain. Anomalies must thus be distinguished from mere outliers, i.e. cases in which the learned model has failed to produce a clear response; it is also necessary to distinguish novel (but meaningful) input from misclassification error within the existing models.We thus apply a methodology of anomaly detection based on comparing the outputs of strong and weak classifiers [8] to the problem of detecting the rule-incongruence involved in the transition from singles to doubles tennis videos. We then demonstrate how the detected anomalies can be used to transfer learning from one (initially known) rule-governed structure to another. Our ultimate aim, building on existing annotation technology, is to construct an adaptive system for court-based sport video annotation.

  • FarajiDavar N, de Campos TE, Windridge D, Kittler J, Christmas W. (2011) 'Domain Adaptation in the Context of Sport Video Action Recognition'. Sierra Nevada, Spain: NIPS 2011 Domain Adaptation Workshop

    Abstract

    We apply domain adaptation to the problem of recognizing common actions between differing court-game sport videos (in particular tennis and badminton games). Actions are characterized in terms of HOG3D features extracted at the bounding box of each detected player, and thus have large intrinsic dimensionality. The techniques evaluated here for domain adaptation are based on estimating linear transformations to adapt the source domain features in order to maximize the similarity between posterior PDFs for each class in the source domain and the expected posterior PDF for each class in the target domain. As such, the problem scales linearly with feature dimensionality, making the video-environment domain adaptation problem tractable on reasonable time scales and resilient to over-fitting. We thus demonstrate that significant performance improvement can be achieved by applying domain adaptation in this context.

  • Huang Q, Cox S, Yan F, deCampos TE, Windridge D, Kittler J, Christmas W. (2011) 'Improved Detection of Ball Hit Events in a Tennis Game Using Multimodal Information'. Volterra, Italy : KTH Computer Science and Communication 11th International Conference on Auditory-Visual Speech Processing (AVSP), Volterra, Italy: International Conference on Auditory-Visual Speech Processing

    Abstract

    We describe a novel framework to detect ball hits in a tennis game by combining audio and visual information. Ball hit detection is a key step in understanding a game such as tennis, but single-mode approaches are not very successful: audio detection suffers from interfering noise and acoustic mismatch, video detection is made difficult by the small size of the ball and the complex background of the surrounding environment. Our goal in this paper is to improve detection performance by focusing on high-level information (rather than low-level features), including the detected audio events, the ball’s trajectory, and inter-event timing information. Visual information supplies coarse detection of the ball-hits events. This information is used as a constraint for audio detection. In addition, useful gains in detection performance can be obtained by using and inter-ballhit timing information, which aids prediction of the next ball hit. This method seems to be very effective in reducing the interference present in low-level features. After applying this method to a women’s doubles tennis game, we obtained improvements in the F-score of about 30% (absolute) for audio detection and about 10% for video detection.

  • De Campos T, Barnard M, Mikolajczyk K, Kittler J, Yan F, Christmas W, Windridge D. (2011) 'An evaluation of bags-of-words and spatio-temporal shapes for action recognition'. 2011 IEEE Workshop on Applications of Computer Vision, WACV 2011, , pp. 344-351.

    Abstract

    Bags-of-visual-Words (BoW) and Spatio-Temporal Shapes (STS) are two very popular approaches for action recognition from video. The former (BoW) is an un-structured global representation of videos which is built using a large set of local features. The latter (STS) uses a single feature located on a region of interest (where the actor is) in the video. Despite the popularity of these methods, no comparison between them has been done. Also, given that BoW and STS differ intrinsically in terms of context inclusion and globality/locality of operation, an appropriate evaluation framework has to be designed carefully. This paper compares these two approaches using four different datasets with varied degree of space-time specificity of the actions and varied relevance of the contextual background. We use the same local feature extraction method and the same classifier for both approaches. Further to BoW and STS, we also evaluated novel variations of BoW constrained in time or space. We observe that the STS approach leads to better results in all datasets whose background is of little relevance to action classification. © 2010 IEEE.

  • Faraji Davar N, de Campos T, Windridge D, Kittler J, Christmas W. (2011) 'Domain Adaptation in the Context of Sport Video Action Recognition'. Domain Adaptation Workshop, in conjunction with NIPS, Sierra Nevada, Spain: NIPS 2011 Domain Adaptation Workshop

    Abstract

    We apply domain adaptation to the problem of recognizing common actions between differing court-game sport videos (in particular tennis and badminton games). Actions are characterized in terms of HOG3D features extracted at the bounding box of each detected player, and thus have large intrinsic dimensionality. The techniques evaluated here for domain adaptation are based on estimating linear transformations to adapt the source domain features in order to maximize the similarity between posterior PDFs for each class in the source domain and the expected posterior PDF for each class in the target domain. As such, the problem scales linearly with feature dimensionality, making the video-environment domain adaptation problem tractable on reasonable time scales and resilient to over-fitting. We thus demonstrate that significant performance improvement can be achieved by applying domain adaptation in this context.

  • Faraji Davar N, deCampos TE, Kittler J, Yan F. (2011) 'Transductive Transfer Learning for Action Recognition in Tennis Games'. 3rd International Workshop on Video Event Categorization, Tagging and Retrieval for Real-World Applications (VECTaR), in conjunction with ICCV,
  • Almajai I, Kittler J, De Campos T, Christmas W, Yan F, Windridge D, Khan A. (2010) 'BALL EVENT RECOGNITION USING HMM FOR AUTOMATIC TENNIS ANNOTATION'. Proceedings of Intl. Conf. on Image Proc.,
  • McAuley JJ, deCampos TE, Caetano TS. (2010) 'Unified graph matching in Euclidean spaces'. San Francisco : IEEE Conference on Computer Vision and Pattern Recognition (CVPR),
  • Khan A, Windridge D, De Campos T, Kittler J, Christmas W. (2010) 'Lattice-based anomaly rectification for sport video annotation'. Proceedings - International Conference on Pattern Recognition, , pp. 4372-4375.

    Abstract

    Anomaly detection has received much attention within the literature as a means of determining, in an unsupervised manner, whether a learning domain has changed in a fundamental way. This may require continuous adaptive learning to be abandoned and a new learning process initiated in the new domain. A related problem is that of anomaly rectification; the adaptation of the existing learning mechanism to the change of domain. As a concrete instantiation of this notion, the current paper investigates a novel lattice-based HMM induction strategy for arbitrary court-game environments. We test (in real and simulated domains) the ability of the method to adapt to a change of rule structures going from tennis singles to tennis doubles. Our long term aim is to build a generic system for transferring game-rule inferences. © 2010 IEEE.

  • McAuley JJ, deCampos T, Csurka G, Perronnin F. (2009) 'Hierarchical Image-Region Labeling via Structured Learning'. London : Proceedings of the British Machine Vision Conference,
  • de Campos TE, Babu BR, Varma M. (2009) 'CHARACTER RECOGNITION IN NATURAL IMAGES'. INSTICC-INST SYST TECHNOLOGIES INFORMATION CONTROL & COMMUNICATION VISAPP 2009: PROCEEDINGS OF THE FOURTH INTERNATIONAL CONFERENCE ON COMPUTER VISION THEORY AND APPLICATIONS, VOL 2, Lisbon, PORTUGAL: 4th International Conference on Computer Vision Theory and Applications, pp. 273-280.
  • Klami A, Saunders C, deCampos T, Kaski S. (2008) 'Can relevance of images be inferred from eye movements?'. Proceedings of the 1st ACM International Conference on Multimedia Information Retrieval, Vancouver, Canada: MIR '08

    Abstract

    Query formulation and efficient navigation through data to reach relevant results are undoubtedly major challenges for image or video retrieval. Queries of good quality are typically not available and the search process needs to rely on relevance feedback given by the user, which makes the search process iterative. Giving explicit relevance feedback is laborious, not always easy, and may even be impossible in ubiquitous computing scenarios. A central question then is: Is it possible to replace or complement scarce explicit feedback with implicit feedback inferred from various sensors not specifically designed for the task? In this paper, we present preliminary results on inferring the relevance of images based on implicit feedback about users' attention, measured using an eye tracking device. It is shown that, in reasonably controlled setups at least, already fairly simple features and classifiers are capable of detecting the relevance based on eye movements alone, without using any explicit feedback.

  • deCampos TE, Murray DW. (2006) 'Regression-based Hand Pose Estimation from Multiple Cameras'. New York : Proceedings of the Conference on Computer Vision and Pattern Recognition,
  • de Campos TE, Cuevas WWM, Murray DW. (2006) 'Directing the attention of a wearable camera by pointing gestures'. IEEE COMPUTER SOC SIBGRAPI 2006: XIX Brazilian Symposium on Computer Graphics and Image Processing, Proceedings, Manaus, BRAZIL: 19th Brazilian Symposium on Computer Graphics and Image Processing (SIBGRAPI 2006), pp. 179-186.
  • Mayol WW, Tordoff BJ, deCampos TE, Davison AJ, Murray DW. (2003) 'Active Vision forWearables'. Eurowearable,
  • Tordoff B, Mayol WW, deCampos TE, Murray DW. (2002) 'Head pose estimation for wearable robot control'. Cardiff, Wales : British Machine Vision Conference (BMVC),
  • deCampos TE, Bloch I, Cesar-Jr RM. (2001) 'Feature Selection Based on Fuzzy Distances Between Clusters: First Results on Simulated Data'. Rio de Janeiro, Brazil : International Conference on Advances in Pattern Recognition (ICAPR),
  • deCampos TE, Feris FS, Cesar-Jr RM. (2000) 'Eigenfaces versus Eigeneyes: First Steps Toward Performance Assessment of Representarions for Face Recognition'. Acapulco, Mexico : MICAI,
  • Feris FS, deCampos TE, Cesar-Jr RM. (2000) 'Detection and Tracking of Facial Features in Video Sequences'. Acapulco, Mexico : MICAI,
  • deCampos TE, Feris FS, Cesar-Jr RM. (2000) 'A Framework for Face Recognition from Video Sequences Using GWN and Eigenfeature Selection'. Atibaia, Brazil : WAICV,
  • deCampos TE, Feris FS, Cesar-Jr RM. (2000) 'Improved Face x Non-Face Discrimination Using Fourier Descriptors Through Feature Selection'. Gramado, Brazil : SIBGRAPI,

Books

  • de Campos TE. (2012) Proceedings of the 4th UK Computer Vision Student Workshop (BMVW). BMVA

Book chapters

  • Almajai I, Yan F, de Campos TE, Khan A, Christmas W, Windridge D, Kittler J. (2012) 'Anomaly Detection and Knowledge Transfer in Automatic Sports Video Annotation'. in Weinshall D, Anemüller J, van Gool L (eds.) Detection and Identification of Rare Audiovisual Cues Springer 384, pp. 109-117.

    Abstract

    A key question in machine perception is how to adaptively build upon existing capabilities so as to permit novel functionalities. Implicit in this are the notions of anomaly detection and learning transfer. A perceptual system must firstly determine at what point the existing learned model ceases to apply, and secondly, what aspects of the existing model can be brought to bear on the newly-defined learning domain. Anomalies must thus be distinguished from mere outliers, i.e. cases in which the learned model has failed to produce a clear response; it is also necessary to distinguish novel (but meaningful) input from misclassification error within the existing models. We thus apply a methodology of anomaly detection based on comparing the outputs of strong and weak classifiers [10] to the problem of detecting the rule-incongruence involved in the transition from singles to doubles tennis videos. We then demonstrate how the detected anomalies can be used to transfer learning from one (initially known) rule-governed structure to another. Our ultimate aim, building on existing annotation technology, is to construct an adaptive system for court-based sport video annotation.

Patents

  • McAuley J, de Campos TE. (2009) Unified graph matching in Euclidean spaces and applications to image comparison and retrieval. USA: Article number 12571630

    Abstract

    A first graph embedded in a Euclidean space is modeled by a globally rigid first model graph that includes all vertices and edges of the first graph and has a preselected maximum clique size. The modeling is configured to maintain the preselected maximum clique size by employing an edge adding process that replicates a vertex of a vertex pair connected by an edge. A mapping between vertices of the first graph and vertices of a second graph is computed by optimizing a mapping between vertices of the first model graph and vertices of the second graph.

  • McAuley J, de Campos TE, Csurka G, Perronnin F. (2009) Consistent hierarchical labeling of image and image regions. USA: Article number 12546948

    Abstract

    Classification of image regions comprises: recursively partitioning an image into a tree of image regions having the image as a tree root and at least one image patch in each leaf image region of the tree, the tree having nodes defined by the image regions and edges defined by pairs of nodes connected by edges of the tree; assigning unary classification potentials to nodes of the tree; assigning pairwise classification potentials to edges of the tree; and labeling the image regions of the tree of image regions based on optimizing an objective function comprising an aggregation of the unary classification potentials and the pairwise classification potentials.

  • de Campos TE, Perronnin F. (2009) Modeling Images as sets of Weighted Features. USA, Europe: Article number 12361235

    Abstract

    An apparatus, method, and computer program product are provided for generating an image representation. The method includes receiving an input digital image, extracting features from the image which are representative of patches of the image, generating weighting factors for the features based on location relevance data for the image, and weighting the extracted features with the weighting factors to form a representation of the image.

  • Jones GR, HAmmett BJ, de Campos TE. (2008) Method of and apparatus for processing image data for display by a multiple-view display device. USA, Japan, UK, Europe: Article number 12523450

    Abstract

    A method is provided for processing image data for display by a multiple-view display device (24) so as to reduce the visibility of undesirable artefacts. Image pixel data are received (20, 21) representing the pixel brightnesses of respective images or sequences of images. The pixel data are processed (22) by applying a unidirectional filter. The processed pixel data for the images may then be interleaved (23) and supplied to the display device (24).

Reports

  • deCampos TE, Csurka G, Perronnin F. (2010) Images as Sets of Locally Weighted Features. Guildford, UK : Article number VSSP-TR-1/2010

Theses and dissertations

  • deCampos T. (2006) 3D Visual Tracking of Articulated Objects and Hands.

Follow this link for my recent papers.
My older papers and my thesis are downloadable from Oxford.
Here is a link to my DBLP entry and here is the link to my MS Academic profile.

Further activities

  • I've lead Surrey's effort to join the PASCAL2 network of excellence and now I'm the local site manager (see my PASCAL profile). This should increase our collaboration with PASCAL members and promote more machine learning activity in this research centre.
  • I'm helping with the organisation of BMVC 2012 and I'm chairing the students' workshop
  • I served as one of the area chairs of VISAPP 2012
  • I'm co-chairing the Computer Vision sessions at the EURO 2012 conference
  • I regularly review papers for journals such as Pattern Recognition, PR Letters, Image and Vision Computing, IEEE TIP, and for conferences such as ICCV, ECCV, CVPR, ACCV, ICPR, CIARP, Sibgrapi and a number of workshops.

Page Owner: td0005
Page Created: Tuesday 21 September 2010 15:33:22 by lb0014
Last Modified: Wednesday 21 September 2016 15:35:51 by ak0039
Expiry Date: Wednesday 21 December 2011 15:30:02
Assembly date: Thu Sep 29 20:55:59 BST 2016
Content ID: 37932
Revision: 6
Community: 1379

Rhythmyx folder: //Sites/surrey.ac.uk/CVSSP/people
Content type: rx:StaffProfile