Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks386 from public data
- 2,603
- Springer Handbook of Robotics2,017
The contents have been restructured to achieve four main objectives: the enlargement of foundational topics for robotics, the enlightenment of design of various types of robotic systems, the extension of the treatment on robots moving in the environment, and the enrichment of advanced robotics applications.
- Robust Object Detection with Interleaved Categorization and Segmentation1,083
A novel method for detecting and localizing objects of a visual category in cluttered real-world scenes that is applicable to a range of different object categories, including both rigid and articulated objects and able to achieve competitive object detection performance from training sets that are between one and two orders of magnitude smaller than those used in comparable systems.
- Combined Object Categorization and Segmentation With An Implicit Shape Model1,032
Results for articulated objects, which show that the proposed method can categorize and segment unfamiliar objects in differ- ent articulations and with widely varying texture patterns, even under significant partial occlusion.
- 991
- A Survey on Neural Network Interpretability967
This survey conducts a comprehensive review of the neural network interpretability research and proposes a novel taxonomy organized along three dimensions: type of engagement (passive vs. active interpretation approaches), the type of explanation, and the focus (from local to global interpretability).
- The Visual Object Tracking VOT2015 Challenge Results902
The Visual Object Tracking challenge 2015, VOT2015, aims at comparing short-term single-object visual trackers that do not apply pre-learned models of object appearance and presents a new VOT 2015 dataset twice as large as in VOT2014 with full annotation of targets by rotated bounding boxes and per-frame attribute.
- The Sixth Visual Object Tracking VOT2018 Challenge Results804
The Visual Object Tracking challenge VOT2018 is the sixth annual tracker benchmarking activity organized by the VOT initiative; results of over eighty trackers are presented; many are state-of-the-art trackers published at major computer vision conferences or in journals in the recent years.
- The Visual Object Tracking VOT2016 Challenge Results776
The Visual Object Tracking challenge VOT2016 goes beyond its predecessors by introducing a new semi-automatic ground truth bounding box annotation methodology and extending the evaluation system with the no-reset experiment.
- A Novel Performance Evaluation Methodology for Single-Target Trackers667
The requirements are the basis of a new evaluation methodology that aims at a simple and easily interpretable tracker comparison and a fully-annotated dataset with per-frame annotations with several visual attributes, which is the largest benchmark to date.
- The Visual Object Tracking VOT2014 Challenge Results606
The Visual Object Tracking challenge 2014, VOT2014, aims at comparing short-term single-object visual trackers that do not apply pre-learned models of object appearance and introduces a new dataset with full annotation of targets by rotated bounding boxes and per-frame attribute.
- 496
- 458
- 390
- The Eighth Visual Object Tracking VOT2020 Challenge Results350
A significant novelty is introduction of a new VOT short-term tracking evaluation methodology, and introduction of segmentation ground truth in the VOT-ST2020 challenge – bounding boxes will no longer be used in theVDT challenges.
- Robust Recognition Using Eigenimages327
This paper presents a new approach which successfully solves problems related to outliers, occlusions, and varying background in appearance-based matching methods using eigenimages using a robust hypothesize-and-test paradigm.
- Segmentation and recovery of superquadrics: computational imaging and vision264
This chapter discusses the development of superquadrics and their Geometric Properties, and some of the applications of Superquadrics found in the real world.
- Segmentation and Recovery of Superquadrics264
This chapter discusses the development of superquadrics and their Geometric Properties, and some of the applications of Superquadrics found in the real world.
- Continual learning: A comparative study on how to defy forgetting in classification tasks.262
This work focuses on task-incremental classification, where tasks arrive in a batch-like fashion, and are delineated by clear boundaries, and studies the influence of model capacity, weight decay and dropout regularization, and the order in which the tasks are presented, to compare methods in terms of required memory, computation time and storage.
- 223
- FS-Net: Fast Shape-based Network for Category-Level 6D Object Pose Estimation with Decoupled Rotation Mechanism221
A fast shape-based network (FS-Net) with efficient category-level feature extraction for 6D pose estimation and a novel decoupled rotation mechanism that employs two decoders to complementarily access the rotation information.
- Prompting for Multi-Modal Tracking215
A novel multi-modal prompt tracker (ProTrack), which can transfer the multi- modal inputs to a single modality by the prompt paradigm, and can achieve high-performance multi- Modal tracking by only altering the inputs, even without any extra training on multi-Modal data.
- Multivariate online kernel density estimation with Gaussian kernels211
This document includes some detailed supplemental derivations used in the bandwidth estimation for the online Kernel Density Estimator which was proposed in the paper \Multivariate Online Kernel D density Estimation with Gaussian Kernels.
- 203
- 191
- Superquadrics for segmenting and modeling range data191
It is shown that superquadrics can directly be recovered from unsegmented data, thus avoiding any presegmentation steps (e.g. in terms of surfaces), and the approach is based on the recover-and-select paradigm.
- SQN: Weakly-Supervised Semantic Segmentation of Large-Scale 3D Point Clouds181
It is observed that, as point clouds are samples of the 3D world, the distribution of points in a local neighborhood is relatively homogeneous, exhibiting strong semantic similarity, and a new weak supervision method is proposed to implicitly augment highly sparse supervision signals.
- Segmentation of range images as the search for geometric parametric models166
A new paradigm for the segmentation of range images into piecewise continuous surfaces is presented through an effective combination of simple component algorithms, which stands in contrast to methods which attempt to solve the problem in a single processing step using sophisticated means.
- Wavelet-Based Dual-Branch Network for Image Demoiréing158
Although designed for image demoireing, WDNet has been applied to two other low-levelvision tasks, outperforming state-of-the-art image deraining and derain-drop methods on the Rain100h and Raindrop800 data sets, respectively.
- 145
- One-shot learning and generation of dexterous grasps for novel objects142
A method for one-shot learning of dexterous grasps and grasp generation for novel objects using an incomplete point cloud from a depth camera and a product of experts, in which experts are of two types.
- 139
- 130
- DepthTrack: Unveiling the Power of RGBD Tracking130
A new RGBD tracking dataset - Depth-Track - is introduced that has twice as many sequences and scene types than in the largest existing dataset, and three times more objects, and the average length of the sequences, number of deformable objects and the number of annotated tracking attributes have been increased.
- The Ninth Visual Object Tracking VOT2021 Challenge Results127
The Visual Object Tracking challenge VOT2021 is the ninth annual tracker benchmarking activity organized by the VOT initiative; results of 71 trackers are presented; many are state-of-the-art trackers published at major computer vision conferences or in journals in recent years.
- Image Demoireing with Learnable Bandpass Filters126
Experimental results on two public datasets show that the novel multiscale bandpass convolutional neural network (MBCNN) outperforms state-of-the-art methods by a large margin.
- 121
- G2L-Net: Global to Local Network for Real-Time 6D Pose Estimation With Embedding Vector Features120
A novel real-time 6D object pose estimation framework, named G2L-Net, which operates on point clouds from RGB-D detection in a divide-and-conquer fashion and achieves state-of-the-art performance in terms of both accuracy and speed.
- An efficient MDL-based construction of RBF networks118
By iteratively combining two procedures, this work achieves a controlled way of training and modifying RBF networks, which balances accuracy, training time, and complexity of the resulting network.
- 107
- Dynamic grasp and trajectory planning for moving objects106
This paper shows how a robot arm can follow and grasp moving objects tracked by a vision system, as is needed when a human hands over an object to the robot during collaborative working.
- 97
- HS-Pose: Hybrid Scope Feature Extraction for Category-level Object Pose Estimation94
A simple network structure, the HS-layer, which extends 3D-GC to extract hybrid scope latent features from point cloud data for category-level object pose estimation tasks and outperforms the state-of-the-art methods by a large margin.
- Region-sequence based six-stream CNN features for general and fine-grained human action recognition in videos93
This work seeks to improve fine-grained action discrimination, while also retaining the ability to perform general action recognition, by reducing feature dimension while also effectively combining appearance and motion information in a unified framework.
- 92
- An Implicit Shape Model for Combined Object Categorization and Segmentation92
This work presents a method for object categorization in real-world scenes that automatically segments the object as a result of the categorization.
- MDL Principle for Robust Vector Quantisation92
This work addresses the problem of finding the optimal number of reference vectors for vector quantisation from the point of view of the Minimum Description Length (MDL) principle, and develops an efficient algorithm for optimising the MDL criterion.
- Collaborative Learning for Hand and Object Reconstruction with Attention-guided Graph Convolution91
This work proposes a collaborative learning strategy where two-branches of deep networks are learning from each other, and achieves beyond state-of-the-art accuracy in 3D pose estimation, as well as recovers dense 3D hand and object shapes.
- 89
- 87
- NTIRE 2021 Challenge on High Dynamic Range Imaging: Dataset, Methods and Results87
This paper reviews the first challenge on high-dynamic range (HDR) imaging that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2021.
- 81
- 81
- 80
- 78
- 77
- Robust localization using an omnidirectional appearance-based subspace model of environment71
This paper discusses a framework for visual self-localization of mobile robots using a parametric model built from panoramic snapshots of the environment, and proposes solutions to the problems related to robustness against occlusions and invariance to the rotation of the sensor.
- 71
- The Thermal Infrared Visual Object Tracking VOT-TIR2016 Challenge Results68
A significant general improvement of results has been observed, which partly compensate for the more difficult sequences introduced into the dataset in VOT-TIR2016.
- Weighted and robust learning of subspace representations68
A weighted version of PCA is presented, which, unlike the standard approach, considers individual pixels and images selectively, depending on the corresponding weights, and a robust PCA method is proposed for obtaining a consistent subspace representation in the presence of outlying pixels in the training images.
- Multiple eigenspaces68
A novel self-organizing framework to construct multiple, low-dimensional eigenspaces from a set of training images, based on the MDL principle, is proposed and the signi8cance of the approach with respect to the recognition rate has been demonstrated.
- Finding optimal neural networks for land use classification67
The authors present a fully automatic and computationally efficient algorithm based on the minimum description length principle (MDL) for optimizing multilayer perceptron (MLP) classifiers and compare their results with a hand-designed MLP and a Gaussian maximum likelihood classifier.
- Model-Based Image Signal Processors via Learnable Dictionaries65
This work proposes an invertible model, capable of bidirectional mapping between RAW and RGB domains, that employs end-to-end learning of rich parameter representations that are free from direct parametric supervision and additionally enable simple and plausible data augmentation.
- 62
- NTIRE 2021 Challenge on Perceptual Image Quality Assessment60
The NTIRE 2021 challenge on perceptual image quality assessment (IQA), held in conjunction with the New Trends in Image Restoration and Enhancement workshop (NTIRE) workshop at CVPR 2021, achieved much better results than existing IQA methods, while the winning method can demonstrate state-of-the-art performance.
- 59
- AIM 2019 Challenge on Image Demoireing: Methods and Results57
This paper reviews the first-ever image demoireing challenge that was part of the Advances in Image Manipulation (AIM) workshop, held in conjunction with ICCV 2019, and focuses on the proposed solutions and their results.
- 57
- A Multi-Hypothesis Approach to Color Constancy56
This work proposes a Bayesian framework that naturally handles color constancy ambiguity via a multi-hypothesis strategy and provides state-of-the-art accuracy on multiple public datasets (up to 11% median angular error improvement) while maintaining real-time execution.
- 55
- Learning a Model-Driven Variational Network for Deformable Image Registration54
VR-Net is proposed, a novel cascaded variational network for unsupervised deformable image registration that outperforms state-of-the-art deep learning methods on registration accuracy, while maintains the fast inference speed of deep learning and the data-efficiency of variational model.
- 54
- 54
- How Computer Vision Can Help in Outdoor Positioning54
Two modes of positioning in a challenging real world scenario are explored: single snapshot based positioning, improved by a novel high-dimensional feature matching method, and continuous positioning enabled by combination of snapshot and incremental positioning.
- 53
- 52
- A Robust PCA Algorithm for Building Representations from Panoramic Images52
A novel robust PCA method for obtaining a consistent subspace representation in the presence of outlying pixels in the training images by treating the outlying points as missing pixels, which enables surveillance and view-based mobile robot localization.
- 48
- Online kernel density estimation for interactive learning48
A Gaussian-kernel-based online kernel density estimation which can be used for applications of online probability density estimation and online learning and allows online adaptation from positive examples as well as from the negative examples.
- NTIRE 2020 Challenge on Image Demoireing: Methods and Results45
This paper reviews the Challenge on Image Demoireing that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2020, and examines the current state-of-the-art in image and burst image demoireing problems.
- A framework for visual-context-aware object detection in still images45
A fully probabilistic framework for visual-context-aware object detection is presented and an in-depth analysis is given discussing the contributions of the individual contextual cues and the limitations of visual context for object detection.
- 44
- Learning to Exploit Stability for 3D Scene Parsing44
A novel architecture for 3D scene parsing named Prim R-CNN, learning to predict bounding boxes as well as their 3D size, translation, and rotation is presented, and it is shown that applying physics supervision on unlabeled real images improves real domain transfer of models training on synthetic data.
- Detection of Human Rights Violations in Images: Can Convolutional Neural Networks Help?43
This work introduces a new, well-sampled human rights-centric dataset called Human Rights Understanding (HRUN), and conducts a rigorous evaluation on a common ground by combining this dataset with different state-of-the-art deep convolutional architectures in order to achieve recognition of human rights violations.
- A local-motion-based probabilistic model for visual tracking42
The experiments show that the proposed local-motion model largely resolves situations when the target is occluded by, or moves in front of, a visually similar object.
- 42
- 41
- On the Importance of Accurate Geometry Data for Dense 3D Vision Tasks41
This paper rigorously shows the significant impact of sensor characteristics on the learned predictions and notice generalisation issues arising from various technologies in everyday household environments, and quantifies the considerable sensor noise impact.
- NTIRE 2022 Challenge on High Dynamic Range Imaging: Methods and Results40
This paper reviews the challenge on constrained high dynamic range (HDR) imaging that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2022, and the competition set-up, datasets, proposed methods and their results.
- End-to-End Learning to Grasp via Sampling From Object Point Clouds40
This letter proposes an end-to-end learning solution to generate 6-DOF parallel-jaw grasps starting from the 3D partial view of the object and shows the effectiveness of L2G as well as its robustness and generalization abilities.
- AIM 2019 Challenge on Image Demoireing: Dataset and Study40
This paper introduces a novel dataset, called LCDMoire, which was created for the first-ever image demoireing challenge that was part of the Advances in Image Manipulation (AIM) workshop, held in conjunction with ICCV 2019.
- 40
- Automatic selection of reference views for image-based scene representations40
This approach is able to generate the scene representation and render images automatically and efficiently even for complex scenes of 3-D objects, since there is no attempt to understand the semantics of images.
- Integrated Commonsense Reasoning and Deep Learning for Transparent Decision Making in Robotics39
The architecture combines the complementary strengths of non-monotonic logical reasoning with incomplete commonsense domain knowledge, deep learning, and inductive learning to provide on-demand explanations of its decisions, the evolution of associated beliefs, and the outcomes of hypothetical actions, in the form of relational descriptions of relevant domain objects, attributes, and actions.
- Learning Hierarchical Compositional Representations of Object Structure38
This paper will focus on what it believes are two central representational design principles, namely a hierarchical organization of categorical representations, more specifically, the principle of hierarchical compositionality, and statistical, bottom-up learning.
- Incremental and robust learning of subspace representations37
An incremental method is presented, which sequentially updates the principal subspace considering weighted influence of individual images as well as individual pixels within an image, resulting in a novel method for incremental, weighted, and robust subspace learning.
- FlexHDR: Modeling Alignment and Exposure Uncertainties for Flexible HDR Imaging36
A strategy that learns to jointly align and assess the alignment and exposure reliability using an HDR-aware, uncertainty-driven attention map that robustly merges the frames into a single high quality HDR image is introduced.
- 36
- PLANNING SEQUENCES OF VIEWS FOR 3-D OBJECT RECOGNITION AND POSE DETERMINATION36
An original procedure is developed which analyzes the transformation of individual clusters under changing viewpoints into several smaller clusters and results in an optimal next-view planning when additional views are necessary to resolve the ambiguities.
- A direct recovery of superquadric models in range images using recover-and-select paradigm36
It is shown that a set of superquadric models can be directly recovered from unsegmented range data, as opposed to methods which attempt the recovery of volumetric models only after the data has been pre-segmented using extensive pre-processing.
- Illumination insensitive recognition using eigenspaces34
The major idea is to incorporate a gradient based filter bank into the eigenspace recognition framework and it is demonstrated that the eigenimage coefficients are invariant to linear filtering (input and eigenimages are filtered with same filters).
- 32
- Distractor-Supported Single Target Tracking in Extremely Cluttered Scenes32
The experimental results show that applying the proposed global dynamic constraint in a feedback loop can improve single target tracker performance, and demonstrate that the overall algorithm significantly outperforms other state-of-the-art single target trackers in highly cluttered scenes.
- 32
- 32
- ExSel++: A general framework to extract parametric models32
A framework for accurate and robust extraction of parametric models of different types is presented and the robust fitting technique, which increases the accuracy of the selected models, may change the classification of data elements which requires the final selection.
- Structural descriptions in human-assisted robot visual learning30
The paper shows how structural descriptions enable relating models for different aspects of one and the same object, and how being able to relate descriptions for visual models and discourse referents enables incremental updating of model descriptions through dialogue.
- Superquadrics and Their Geometric Properties30
This chapter defines superquadrics after a brief history of their development is outlined and basic superquadric equations are given.
- Panoramic Eigenimages for Spatial Localisation30
This paper presents a pure appearance-based localisation method using an eigenspace representation of panoramic images and evaluates simple localisation of images not included in the training set.
- Transformer-Based Models for Predicting Molecular Structures from Infrared Spectra Using Patch-Based Self-Attention29
A transformer-based model with a patch-based self-attention spectrum embedding layer, designed to prevent the loss of spectral information while maintaining simplicity and effectiveness, achieves state-of-the-art performance on simulated data sets and demonstrates proficiency in analyzing intricate and variable fingerprint regions.
- Answering Visual What-If Questions: From Actions to Predicted Scene Descriptions29
This paper proposes a hybrid model which integrates a physics engine into a question answering architecture in order to anticipate future scene states resulting from object-object interactions caused by an action.
- Agri-LLaVA: Knowledge-Infused Large Multimodal Assistant on Agricultural Pests and Diseases28
The first multimodal instruction-following dataset in the agricultural domain is constructed, covering over 221 types of pests and diseases with approximately 400,000 data entries, and a knowledge-infused training method is proposed to develop Agri-LLaVA, an agricultural multimodal conversation system.
- Unsupervised Model Personalization While Preserving Privacy and Scalability: An Open Problem28
This work investigates the task of unsupervised model personalization, adapted to continually evolving, unlabeled local user images, and provides a novel Dual User-Adaptation framework (DUA) to explore the problem.
- A Coarse-to-Fine Taxonomy of Constellations for Fast Multi-class Object Detection28
A novel approach for speeding up recognition times of multi-class part-based object representations by constructing a taxonomy of constellation models cascaded from coarse-to-fine resolution and using it in recognition with an efficient search strategy.
- 27
- 27
- 27
- Deep neural networks and image classification in biological vision26
It is shown that there are important divergences between fDCNNs and biological vision that reflect fundamental differences in computational architectures, and representational structures, supporting image classification in these networks and the biological system.
- 26
- 26
- 25
- 25
- Selecting features for object detection using an AdaBoost-compatible evaluation function24
This paper addresses the problem of selecting features in a visual object detection setup where a detection algorithm is applied to an input image represented by a set of features using a novel evaluation function that measures the utility of individual features for a certain detection task.
- Why to combine reconstructive and discriminative information for incremental subspace learning24
A novel method for in- cremental visual learning by combining reconstructive and discriminative subspace methods that is able to efficiently update the current model with new instances of the already learned classes as well as to introduce new classes.
- Is my Depth Ground-Truth Good Enough? HAMMER -- Highly Accurate Multi-Modal Dataset for DEnse 3D Scene Regression23
HAMMER is proposed, a dataset comprising depth estimates from multiple commonly used sensors for indoor depth estimation, namely ToF, stereo, active stereo together with monocular RGB+P data that provides a reliable base to pave the way to targeted depth improvements and sensor fusion approaches.
- Wild ToFu: Improving Range and Quality of Indirect Time-of-Flight Depth with RGB Fusion in Challenging Environments23
A new learning based end-to-end depth prediction network which takes noisy raw I-ToF signals as well as an RGB image and fuses their latent representation based on a multi step approach involving both implicit and explicit alignment to predict a high quality long range depth map aligned to the RGB viewpoint is proposed.
- Beyond Standard Benchmarks: Parameterizing Performance Evaluation in Visual Object Tracking23
This paper presents an approach that utilizes omnidirectional videos to generate realistic, consistently annotated, short-term tracking scenarios with exactly parameterized motion patterns, and provides an in-depth analysis of major tracking paradigms which is complementary to the standard benchmarks.
- 23
- 23
- Color image segmentation and color constancy22
An approach to the construction of computational model for color image segmentation based on the physical properties of sensors, illumination lights and surface reflectances, and develops a metric space of intensity, hue and saturation in order to better interpret light- surface interactions.
- 22
- The VOT2013 challenge: overview and additional results21
An overview of the VOT2013 challenge is provided, its main results are pointed out, and the additional previously unpublished experiments and results are documented.
- 21
- Evaluating Deep Convolutional Neural Networks for Material Classification20
A rigorous evaluation of how state-of-the art CNN architectures compare on a common ground over widely used material databases shows that the best performing CNN architectures can achieve up to 94.99\% mean average precision when classifying materials.
- 20
- 20
- 20
- Wavelet-based network for high dynamic range imaging19
A novel frequency-guided end-to-end deep neural network (FHDRNet) is proposed to conduct HDR fusion in the frequency domain, and Discrete Wavelet Transform (DWT) is used to decompose inputs into different frequency bands.
- CroMo: Cross-Modal Learning for Monocular Depth Estimation19
This paper proposes a novel pipeline capable of estimating depth from monocular polarisation, and proposes the inversion of differentiable analytic models thereby connects scene geometry with polarisation and ToF signals and enables self-supervised and cross-modal learning.
- 19
- 19
- Planning the next view using the max-min principle19
This paper presented an approach to a typical task of active sensing, namely how to successively position the sensor in order to accomplish a given task, and applied it to the task of determining the 3-D coordinates of the points of object silhouettes in the image under orthographic projection.
- 18
- SQN: Weakly-Supervised Semantic Segmentation of Large-Scale 3D Point Clouds with 1000x Fewer Labels.18
- Towards deep compositional networks18
This work proposes a novel analytic model of a basic unit in a layered hierarchical model with both explicit compositional structure and a well-defined discriminative cost function, affording a straight-forward visualization of parts and faster inference due to separability of the units.
- Evaluating multi-class learning strategies in a generative hierarchical framework for object detection18
This paper shows how different types of multi-class learning can be done within one generative hierarchical framework and provides a rigorous experimental analysis of various object class learning strategies as the number of classes grows.
- Finding parametric curves in an image18
The presented method is efficient for two reasons: firstly, it is designed as a search which utilizes intermediate results as a guidance toward the final result, and secondly, it combines model recovery and model selection in a computationally efficient procedure.
- 17
- ILSH: The Imperial Light-Stage Head Dataset for Human Head View Synthesis17
The setup of a light-stage specifically designed to capture high-resolution (4K) human head images and the process of addressing challenges (preprocessing, ethical issues) in collecting high-quality data are described.
- Repurposing existing deep networks for caption and aesthetic-guided image cropping17
Through various quantitative and qualitative evaluations, it is shown that the proposed novel optimization framework can produce crops that are well-aligned to intended user descriptions and aesthetically pleasing.
- RGBD Object Tracking: An In-depth Review17
This paper is the first to provide depth quality evaluation and analysis of tracking results in depth-friendly scenarios in RGBD tracking, and proposes robustness evaluation against input perturbations.
- 17
- 17
- UNSUPERVISED LEARNING OF A HIERARCHY OF TOPOLOGICAL MAPS USING OMNIDIRECTIONAL IMAGES17
This paper presents a novel appearance-based method for path-based map learning by a mobile robot equipped with an omnidirectional camera that provides an abstraction of the environment in terms of visual aspects and demonstrates how the resulting representation can be used for modeling indoor and outdoor environments.
- A System for Learning Basic Object Affordances using a Self-Organizing Map17
The development of a cognitive vision platform that uses a robotic arm to interact with household objects in an attempt to learn some of their basic affordance properties is discussed and the various sensor and effector module competencies that were needed are outlined.
- 16
- Cognitive Systems16
It is argued that there is no need to attribute to such systems as a whole any form of cognitive agency, and that the authors can well understand the importance of such systems while restricting agency to the human components.
- 16
- Towards correct and informative evaluation methodology for texture classification under varying viewpoint and illumination16
This work accurately registers the images across all imaging conditions and split the surface patches to parts and shows that such methodology gives a more realistic assessment of classifier performance.
- CONTINUOUS LEARNING OF SIMPLE VISUAL CONCEPTS USING INCREMENTAL KERNEL DENSITY ESTIMATION16
This paper proposes a method for continuous learning of simple visual concepts that continuously associates words describing observed scenes with automatically extracted visual features with kernel density probability distribution estimates.
- Context Driven Focus of Attention for Object Detection16
This work presents a concept of how to extract and learn contextual information from examples that represents a prior for object detection, and shows that context awareness provides complementary information over pure local appearance-based processing.
- Conservative Visual Learning for Object Detection with Minimal Hand Labeling Effort16
By using this approach to start with a very simple object detection system and using generative and discriminative classifiers in an iterative co-training fashion to arrive at increasingly better object detectors, the framework achieves a state of the art detection rate.
- 15
- HDR Reconstruction from Bracketed Exposures and Events15
This paper presents a multi-modal end-to-end learning-based HDR imaging system that fuses bracketed images and event modalities in the feature domain using attention and multi-scale spatial alignment modules and proposes a novel event- to-image feature distillation module that learns to translate event features into the image-feature space with self-supervision.
- 15
- 15
- GeoReF: Geometric Alignment Across Shape Variation for Category-level Object Pose Refinement14
This work introduces a novel architecture for category-level object pose refinement and introduces a cross-cloud transformation mechanism that efficiently merges di-verse data sources and pushes the limits of the model by incorporating the shape prior information for translation and size error prediction.
- Exploring Object-Centric and Scene-Centric CNN Features and Their Complementarity for Human Rights Violations Recognition in Images14
The human rights archive (HRA) database is introduced, a verified-by-experts repository of 3050 human rights violations photographs, labeled with human rights semantic categories, comprising a list of the types of human rights abuses encountered at present.
- Robust Real-Time Music Transcription with a Compositional Hierarchical Model14
A new compositional hierarchical model for robust music transcription whose main features are unsupervised learning of a hierarchical representation of input data, transparency, which enables insights into the learned representation, as well as robustness and speed which make it suitable for real-world and real-time use.
- 14
- Proper scale for modeling visual data14
A method for determining the proper scale for modeling visual data and an efficient architecture for selective image modeling is discussed which selects models according to the task, the nature of the scene and the computational constraints.
- Many-Shot from Low-Shot: Learning to Annotate Using Mixed Supervision for Object Detection13
This work introduces, for the first time, an online annotation module (OAM) that learns to generate a many-shot set of \emph{reliable} annotations from a larger volume of weakly labelled images, resulting in a fully end-to-end strategy that only requires a low- shot set of fully annotated images.
- SymCHM—An Unsupervised Approach for Pattern Discovery in Symbolic Music with a Compositional Hierarchical Model13
A compositional hierarchical model for pattern discovery in symbolic music that can learn a set of repeated patterns within individual works or larger corpora in an unsupervised manner, relying on statistics of pattern occurrences, and robustly infer the learned patterns in new, unknown works is presented.
- Learning Hierarchical Representations of Object Categories for Robot Vision13
This paper presents the recently developed approach to constructing a hierarchical representation of visual input that aims to enable recognition and detection of a large number of object categories by learning a hierarchy of spatially flexible compositions in an unsupervised, statistics-driven manner.
- 12
- Spatially-Adaptive Filter Units for Deep Neural Networks12
This paper proposes a novel displaced aggregation unit (DAU) that does not require hand-crafting, and extensively demonstrate the strength of DAUs on a classification and semantic segmentation tasks.
- 12
- Multi-task Learning with 3D-Aware Regularization11
This work proposes to address the problem of cross-task correlations in deep neural networks by introducing a structured 3D-aware regularizer which interfaces multiple tasks through the projection of features extracted from an image encoder to a shared 3D feature space and decodes them into their task output space through differentiable rendering.
- 11
- An Analysis of Rhythmic Patterns with Unsupervised Learning11
A model capable of learning distinctive rhythmic structures of different music genres using unsupervised learning that employs tempo-invariant encoding of patterns and can thus learn and perform inference on tempo-varying and noisy input data is presented.
- 11
- Evaluating multi-class learning strategies in a hierarchical framework for object detection11
It is shown that sequential training achieves the best trade-off between inference and training times at a comparable detection performance and could thus be used to learn the classes on a larger scale.
- Optimization framework for learning a hierarchical shape vocabulary for object class detection11
The idea of this paper is to cast the vocabulary learning into an optimization framework that iteratively improves the hierarchy as a whole and attains a competitive detection performance with respect to the current state-of-the-art at both, faster inference as well as shorter training times.
- 11
- A Framework for Robust and Incremental Self-Localization of a Mobile Robot11
A framework for an embodied robotic system that is capable of appearance-based self-localization based on a panoramic eigenspace model of the environment and allows for simultaneous localization and map building using an incremental learning algorithm is presented.
- A Framework for Robust and Incremental Self-Localization11
A framework for an embodied robotic system that is capable of appearance-based self-localization based on a panoramic eigenspace model of the environment and allows for simultaneous localization and map building using an incremental learning algorithm is presented.
- LEE: A Photorealistic Virtual Environment for Assessing Driver-Vehicle Interactions in Self-driving Mode10
A low-cost, extendable, and easy-to-use virtual environment with the aim to highlight its utility for level 3 driving automation and shows that the speed of the car at the time when the system needs to transfer control to the human driver is critical.
- 10
- 10
- 10
- Two-dimensional object recognition using multiresolution non-information-preserving shape features10
A set of non-information-preserving features based on a multiresolution curve analysis which makes shape features explicit at multiple scales is proposed.
- Dynamic multi-level appearance models and adaptive clustered decision trees for single target tracking9
This paper presents a tracking algorithm for arbitrary objects in challenging video sequences that is more robust than the state-of-the-art methods from both of those tracking challenges, while also offering competitive tracking precision.
- Image-based Registration for a Neurosurgical Robot: Comparison Using Iterative Closest Point and Coherent Point Drift Algorithms9
This work compares two registration algorithms, iterative closest point (ICP) and coherent point drift (CPD), by registering ideal point clouds taken from MRI data with re-meshed, noisy and smoothed versions and finds that ICP generally gives better and more consistent registration accuracy for the region of interest than CPD.
- A local-global coupled-layer puppet model for robust online human pose tracking9
A novel cross-coupled global-local model of articulated human body pose is proposed, which uses the previous notions of deformable structure (DS) puppet models, and an adaptive penalty function is proposed for optimizing the pose estimates.
- Efficient View Synthesis and 3D-based Multi-Frame Denoising with Multiplane Feature Representations8
This work introduces the first 3D- based multi-frame denoising method that significantly outperforms its 2D-based counterparts with lower computational requirements.
- Residual Contrastive Learning for Image Reconstruction: Learning Transferable Representations from Noisy Images8
A new label-efficient learning paradigm based on residuals, residual contrastive learning (RCL), is proposed, and an unsupervised visual representation learning framework is derived, suitable for low-level vision tasks with noisy inputs.
- 8
- Room Categorization Based on a Hierarchical Representation of Space8
This paper proposes a new compositional hierarchical representation of space that is based on learning statistically significant observations, in terms of the frequency of occurrence of various shapes in the environment, on a two-dimensional space.
- 8
- 8
- Articulation in Motion: Prior-free Part Mobility Analysis for Articulated Objects By Dynamic-Static Disentanglement7
A new framework for articulation in Motion is presented, which infer part-level decomposition, articulation kinematics, and reconstruct an interactive 3D digital replica from a user-object interaction video and a start-state scan, and proposes a dual-Gaussian scene representation that uses motion cues to segment the object into parts and assign articulation joints.
- Label-efficient object detection via region proposal network pre-training7
This work proposes a simple pretext task that provides an effective pre-training for the RPN, towards efficiently improving downstream object detection performance, and evaluates the efficacy of the approach on benchmark object detection tasks and additional downstream tasks, including instance segmentation and few-shot detection.
- 7
- Towards Generic 3D Tracking in RGBD Videos: Benchmark and Baseline7
This paper investigates a novel problem: is it possible to track a generic 3D object in RGBD videos and predict 3D bounding boxes of the object of interest and proposes an effective tracking baseline to estimate 3D bounding boxes for arbitrary objects in RGBD videos, by fusing appearance and spatial information effectively.
- 7
- Recognizing 2-tone images in grey-level parametric eigenspaces7
A novel approach is proposed which demonstrates that having an eigenspace encompassing a set of grey-level images, it is possible to recover the eIGenspace coefficients and consequently recognize the input images even in the cases when theinput images are 2-tone images of the objects in the training set.
- 7
- 7
- 6
- Self-supervised HDR Imaging from Motion and Exposure Cues6
Experimental results show that the HDR models trained using the proposed self-supervision approach achieve performance competitive with those trained under full supervision, and are to a large extent superior to previous methods that equally do not require any supervision.
- 6
- Performance Characterization of Image Feature Detectors in Relation to the Scene Content Utilizing a Large Image Database6
This paper presents a new framework for determining the type of scenes which maximize and minimize the performance of detectors in terms of repeatability rate and provides new insights into the behavior of feature detectors.
- Adding discriminative power to a generative hierarchical compositional model using histograms of compositions6
The generative model of hierarchical compositions is combined with the discriminative descriptor by performing hypothesis verification of detections produced by the hierarchical compositional model and is compared against a state-of-the-art convolutional neural network and shown to outperform it under significant occlusions.
- Editorial: Hierarchical Object Representations in the Visual Cortex and Computer Vision6
Over the past 40 years, Neurobiology and Computational Neuroscience have proved that deeper understanding of visual processes in humans and non-human primates can lead to important advancements in computational perception theories and systems.
- 6
- Hardware Implementation of FAST Algorithm for Mobile Applications6
Hardware implementation of FAST algorithm, capable of processing images at constant speed of one pixel per clock is presented, showing that nearly forty times faster corner detection could be achieved on mobile object detection and localization application, if the existing software detector is replaced by the hardware module.
- Towards Learning Basic Object Affordances from Object Properties6
A system that uses a robotic arm, camera systems and self-organizing maps to learn basic affordances of objects to recognize and exploit environmental affordances is presented.
- 6
- Computer vision -- ECCV 2006 : 9th European Conference on Computer Vision, Graz, Austria, May 7-13, 2006 : proceedings6
This work presents novel Unsupervised, Automatic and Robust Methods of Segmentation of Nuclei and Probes, and a simple solution to the Six-Point Two-View Focal-Length Problem.
- 6
- 6
- 6
- 6
- 5
- BURG-Toolkit: Robot Grasping Experiments in Simulation and the Real World5
BURG-Toolkit, a set of open- source tools for Benchmarking and Understanding Robotic Grasping, is presented and proof-of-concept experimental results quantifying the sim-to-real gap for robot grasping in some example scenes are provided.
- Content-Diverse Comparisons improve IQA5
Evaluation on multiple benchmarks, covering a wide range of distortions and image content, shows the effectiveness of the learning scheme for training image quality assessment models, which includes differentiable regularizers, derived from correlation coefficients, to better adjust predicted scores relative to one another.
- Depth-only Object Tracking5
The end-to-end trained RGBD-DiMP outperforms the recent VOT 2020 RGBD winners and generates depth maps for tracking from the scratch with the generated data and fine-tune it with the available small RGBD tracking datasets.
- 5
- A Generic Framework for Assessing the Performance Bounds of Image Feature Detectors5
A generic framework based on the repeatability measure is presented that allows assessment of the upper and lower bounds of detector performance and finds statistically significant performance differences between detectors as a function of image transformation amount by introducing a new variant of McNemar’s test in an effort to design more reliable and effective vision systems.
- 5
- Data-driven human grasp movement analysis5
A methodology to reconcile two distinct and apparently incompatible approaches to grasping data dimensionality reduction under a unified framework is presented, which allows for a data-generated grasp taxonomy along with low-dimensional representations which could be used for human grasping data classification and posture reconstruction, as well as for simplifying grasp planning algorithms and robotic hands programming.
- Modeling binding and cross-modal learning in Markov logic networks5
By these definitions a binding mechanism in a Markov logic network is model and its role in a cognitive architecture is defined and a prototype binding system is evaluated off-line, using three different inference methods.
- Hyperlinking reality via camera phones5
A novel user interface concept for camera phones, called “Hyperlinking Reality via Camera Phones”, provides a solution to one of the main challenges facing mobile user interfaces, that is, the problem of selection and visualization of actions that are relevant to the user in her current context.
- Integrating Visual Context and Object Detection within a Probabilistic Framework5
The prior probability function of an object detector is applied, that maps the detector's output to probabilities, together with an appropriate contextual weighting, a probabilistic framework is established and results show that visual context facilitates object detection methods.
- 5
- A New Refinement Method for Registration of Range Images Based on Segmented Data5
A new method for registration of range images, which is based on the results the authors obtain from the segmentation process, and which has shown a significant improvement in precision of the registration in comparison with traditional approches.
- 5
- A Robust Approach to Estimation of Parametric Models5
A robust method for estimation of parametric models with a high degree of resistance to outliers and the insensitivity to incorrect initial estimates, successfully applied to linear as well as nonlinear parameter estimation problems.
- Recognition and pose determination of 3-D objects using multiple views5
This work presents a method for automatic recognition and pose (orientation) determination of 3-D objects of arbitrary shape that is tested on a set of real objects using multiresolution NIP (non-information-preserving) shape features to characterize the views.
- Segmentation of Range Images as the Search for the Best Description of the Scene in Terms of Geometric Primitives5
A new paradigm for the segmentation of range images into piecewise continuous patches is presented and a fast optimization procedure for model selection is developed, in combining model extraction and model selection in a dynamic way.
- Improving Object Detection via Local-global Contrastive Learning4
This work presents a novel image-to-image translation method that optimises the appearance of object instances through spatial attention masks, implicitly delineating the scene into foreground regions associated with the target object instances and background non-object regions.
- NCRF: Neural Contact Radiance Fields for Free-Viewpoint Rendering of Hand-Object Interaction4
This work presents a novel free-viewpoint rendering framework, Neural Contact Radiance Field (NCRF), to reconstruct hand-object interactions from a sparse set of videos, and shows that this approach outperforms the current state-of-the-art in terms of both rendering quality and pose estimation accuracy.
- Tunable Convolutions with Parametric Multi-Loss Optimization4
The key insight is to use a shared set of parameters to dynamically interpolate both the objectives and the kernels, which includes an equal number of objectives, to optimize a parametric tunable convolutional layer which includes a number of different kernels.
- G-DAIC: A Gaze Initialized Framework for Description and Aesthetic-Based Image Cropping4
The runtime limitation of the state-of-the-art method is addressed by implementing the Early termination strategy to reduce the number of iterations required to produce the output and the adaptive Mixed scaling method to find the optimal output despite the size of the generated initialisation region and the described part of the image.
- Material Classification in theWild: Do Synthesized Training Data Generalise Better than Real-world Training Data?4
It is demonstrated that synthesized data achieve an improvement on mean average precision when used as training data and in conjunction with pre-trained CNN architectures, which spans from ~ 5% to ~ 19% across three widely used material databases of real-world images.
- 4
- A Framework for Continuous Learning of Simple Visual Concepts4
A framework for learning sim-ple visual concepts that addresses the problem of cross-modal learning of elementary visual properties and spatial relations and shows that the same learning mechanism can be used to learn both types of concepts.
- 4
- 4
- 4
- 4
- View-based object representations using RBF networks4
This paper proposes a systematic approach for building object representations in terms of RBF networks, and studied and designed two procedures: the off-line procedure, where the network is constructed after having a complete set of training views of an object, and the on-line Procedure, whereThe network is incrementally built as new views of a object arrive.
- Reverse Engineering by Means of Range Image Interpretation4
The approach to segmentation and modeling of range images which enables the reconstruction of a compact geometric representation in the MDL sense which may consist of different types of models.
- Using recover-and-select paradigm on triangulated data4
A new approach to segment triangulated data and to extract geometric models out of it using Recover-and-select paradigm, which successfully accomplishes these tasks when applied to range images and can be applied also to triangulation data.
- Stereo matching using M-estimators4
The standard area-based correlation approach is modified so that it can tolerate a significant number of outliers and exhibits a robust behaviour not only in the presence of mismatches but also in the case of depth discontinuities.
- Computational Aspects of Color Constancy4
It is shown that, within finitedimensional models, formulation and solution of color constancy are determined by the choice of basis functions, the number of spectral receptors and the spatial constraints.
- 3
- 3
- Disentangling 3D Attributes from a Single 2D Image: Human Pose, Shape and Garment3
The method is the first to achieve cross-domain disentanglement for this highly under-constrained problem of extracting disentangled 3D attributes only from 2D image data and demonstrates the framework's ability to transfer pose, shape, and garments in 3D reconstruction on virtual data.
- Motion-aware ensemble of three-mode trackers for unmanned aerial vehicles3
The proposed motion-aware ensemble shows a 5.3% improvement in average precision compared to the baseline correlation filter tracker, which effectively employs deep features while achieving a tracking speed of at least 80fps in the authors' experimental settings.
- 3
- Assessing Capsule Networks with Biased Data3
This paper assesses dynamic routing and EM routing based Capsule Networks and proposes a comparison with Convolutional Neural Networks in the two tested scenarios and provides new insights into the behaviour of capsule networks.
- Motor memory: Representation, learning and consolidation3
A new motor memory architecture is proposed that enables information sharing between different skills, on-line learning and off-line memory consolidation and the integration of motor memory with sensor data into a common sensorimotor memory.
- 3
- Adding Discriminative Power to Hierarchical Compositional Models for Object Class Detection3
This paper proposes a novel approach that preserves the appealing properties of the generative hierarchical models, while at the same time improves their discrimination properties by introducing a network of discriminative nodes on top of the existing generative hierarchy.
- 3
- 3
- Recognizing Objects by Their Appearance Using Eigenimages3
The basic methods for appearance-based object recognition are reviewed and the major limitations of the standard approach are identified and how these limitations can be alleviated leading to an object recognition system which is applicable in real world situations.
- Acquiring Range Images of Objects with Non-Uniform Reflectance Using High Dynamic Scale Radiance Maps3
It is demonstrated that range images obtained from high dynamic scale radiance maps are of much better quality than those obtained directly from the original images of a limited dynamic scale.
- Registration of Range Images Based on Segmented Data3
A new method for registration of range images, which is based on the results the authors obtain from the segmentation process, has shown a significant improvement in precision of the registration in comparison with traditional approach.
- Vector Quantization and Minimum Description Length3
This paper forms the VQ in terms of the MDL principle, and derives depending on the coding procedure different instantiations of the algorithm, and develops an efficient algorithm (similar to EM-type algorithms) for optimizing theMDL criterion.
- 3
- FROM EDGELS TO PARAMETRIC CURVES3
This paper presents a robust algorithm which is designed to work concurrently with diierent curve types where each curve type selects its own domain of applicability.
- Planning the optimal set of views using the Max-Min principle3
The approach is data driven since it is only the data that provides the information for planning the next views, and no a priori knowledge about the constituents of the scene is necessary.
- 2
- Learning part-based spatial models for laser-vision-based room categorization2
A new approach for room categorization that is based on two-dimensional laser range data and a new discriminative dictionary learning technique that is applied for part-dictionary selection in both laser-based and vision-based modalities are presented.
- Visual Object Class Recognition2
This chapter is mostly concerned with the recognition and detection of basic level object classes such as cars, persons, chairs, or dogs.
- 2
- 2
- Learning Compositional Hierarchies of a Sensorimotor System2
It is shown that such a representation enables efficient learning and inference of robot states and relates to a recent research in cognitive science, which provides a solid evidence that perception and action share common neural architectures.
- A Web-Service for Object Detection Using Hierarchical Models2
This paper proposes an architecture for an object detection system suitable for a web-service running distributed on a cluster of machines and provides implementation details for running hierarchical models on top of a distributed platform and proposes an additional hypothesis verification step to reduce many false-positives that are common in hierarchical models.
- 2
- A basic cognitive system for interactive learning of simple visual concepts2
In this work, a system and underlying representations and mechanisms for continuous learning of visual concepts in dialogue with a human tutor are presented.
- 2
- 2
- Editorial Special Issue ECCV 20062
The special issue presents six considerably extended versions of the papers selected by a two-stage process from the 811 submissions to the 9th European Conference on Computer Vision (ECCV 2006), the premium conference on computer vision, held in Graz, Austria, in May 2006.
- Multiresolution Approach to Biomedical Image Segmentation with Statistical Models of Appearance2
It is shown that biomedical image segmentation with statistical models of appearance can be improved in terms of accuracy and efficiency by a multiresolution approach.
- Untitled2
This work proposes two different algorithms for solving the optimization problem of selecting the optimal set of reference images from a large set of images representing the scene and shows the feasibility of the approach for one-parameter camera motion.
- 2
- 1
- 1
- 1
- Cognitive Relevance Transform for Population Re-Targeting1
The results of the experiments have shown that the target population preferred such a transformed categorization by a large margin, that the performance of human observers is probably better than previously thought, and that the outcome of re-targeting may be difficult to predict without actual tests on thetarget population.
- Towards Categorization and Pose Estimation of Sets of Occluded Objects in Cluttered Scenes from Depth Data and Generic Object Models Using Joint Parsing1
A novel multi-stage algorithm is proposed to perform the task of categorizing and estimating the six-dimensional poses of all visible and partly occluded objects present in a scene from depth image information, in the absence of ground truth training examples and exact geometrical models of objects.
- Automatic Selection of the Optimal Local Feature Detector1
A tool capable of quickly analysing input images to determine which type and amount of transformation is applied to them and then selecting the optimal feature detector, which is expected to perform the best is presented.
- Categorisation of 3D Objects in Range Images Using Compositional Hierarchies of Parts Based on MDL and Entropy Selection Criteria1
Two methods based on Minimum Description Length and category based entropy are introduced in a new approach to object categorisation in range images using a novel hierarchical compositional representation of surfaces.
- Hierarchical Compositional Representations of Object Structure1
This talk proposes an approach for speeding up recognition times of multi-class part-based object representations by constructing a taxonomy of constellation models cascaded from coarse-to-fine resolution and use it in recognition with an efficient search strategy.
- 1
- 1
- 1
- 1
- 1
- 1
- Applications of Superquadrics1
The earlier chapters have already shed some light on the utility of superquadrics in computer vision, but it is important to understand the limitations of the basic superquadric models, namely their limited shape vocabulary, and the fact that they are really coarse-grain closed volumetric models suitable, in particular, for object-centered generic shape descriptions.
- 1
- Computer Analysis of Images and Patterns: 8th International Conference, CAIP'99 Ljubljana, Slovenia, September 1-3, 1999 Proceedings1
Object Recognition with Representations Based on Sparsified Gabor Wavelets Used as Local Line Detectors and Effective Implementation of Linear Discriminant Analysis for Face Recognition and Verification.
- 1
- Optimising Neural Networks for Land Use Classification1
A fully automatic and computationally efficient algorithm for optimising multilayer perceptron classifiers that produces a better classification accuracy with a smaller number of hidden units than the hand designed network.
- Planning multiple views for 3-D object recognition and pose determination1
An original procedure is developed which analyzes the transformation of the individual clusters under changing viewpoints into several smaller clusters and results in an optimal next-view planning when additional views are necessary to resolve the ambiguities.
- Agora: Enhancing LLM Agent Reasoning Via Auction-Based Task Allocation–
Agora is a framework that uses a confidence-calibrated auction to dynamically allocate tasks to expert models and tools, and treats reasoning steps as tradeable items, which bases allocation on calibrated competence rather than raw confidence.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- A Multi-Hypothesis Classification Approach to Color Constancy–
This work proposes a Bayesian framework that naturally handles color constancy ambiguity via a multi-hypothesis strategy and provides state-of-the-art accuracy on multiple public datasets (up to 11% median angular error improvement) while maintaining real-time execution.
- –
- Vision for Autonomous Vehicles and Probes (Dagstuhl Seminar 15461)–
This work strictly focuses on mathematical, geometrical and computational aspects of autonomous vehicles and autonomous vehicular technology which make use of computer vision and pattern recognition as the central component for autonomous driving and navigation and remote exploration.
- –
- –
- –
- –
- Binding and Cross-Modal Learning in Markov Logic Networks–
The problems of high-level binding and crossmodal learning are defined and a binding mechanism and a cross-modal learner are model and tested in a Markov logic network and on a synthetic object database.
- –
- –
- –
- Computer Vision -- ECCV 2006 (vol. # 3953) : 9th European Conference on Computer Vision, Graz, Austria, May 7-13, 2006, Proceedings, Part III–
These are the proceedings of the 9th European Conference on Computer Vision (ECCV 2006), the premium European conference on computer vision, held in Graz, Austria, in May 2006, where 811 papers were received, the largest number of submissions so far.
- –
- Activity Recognition via Autoregressive Prediction of Velocity Distribution–
A novel approach for view-based learning and recognition of motion patterns of articulated objects by forming the intervals of motion as a predictive model of local spatio-temporal receptive field activation using the MDL principle.
- –
- –
- –
- –
- –
- –
- –
- Segmentation with Superquadrics–
A common underlying task of most recognition applications is building the scene description in terms of symbolic entities, where each piece of information must be mapped either to a shape primitive or discarded as noise.
- Extensions of Superquadrics–
Hyperquadrics, which include superquadrics as a special case, and are generated by taking hyperslices of high-dimensional algebraic hypersurfaces, are also described at the end of the chapter, as well as ratioquadrics which are very similar to super quadrics but have continuous first derivatives everywhere on the surface.
- Recovery of Individual Superquadrics–
One of the criteria for selection of geometric primitives in computer vision is their accessibility (Brady, 1983) and these primitives should have local support, so that they can cope with occlusions and self-occlusions.
- –
- On-line Learning of Object Representations–
A systematic approach for building object representations in terms of RBF networks is proposed and two procedures are studied and designed: the “off-line” procedure, where the network is constructed after having a complete set of training views of an object, and the "on-line" procedure, which is incrementally built as new views of a object arrive.
- Computational complexity reduction in eigenspace approaches–
This paper proposes to use a focus-of-attention (FOA) detector which is intended to select candidate areas of interest with minimal computational effort and employs a multiresolution approach to alleviate the problems related to complexity.
- –
- –
- –
- Choosing reference views for image-based representation–
Two different algorithms solving the problem, how to select the optimal set of reference views from a dense set of real primary views are proposed, posed as a selection and fitting of parametric models.
- –
- Object-oriented analysis and design of image segmentation package–
An example of object oriented analysis and design of an image segmentation package is presented to show how programming languages and design techniques not only allow us to implement the authors' ideas but also provide means to reene their sloppy concepts.
- –
- –
- –
- –
- –
- Shape modeling through parametric descriptions–
The approach to shape modeling is based on parametric description of all aspects of shape: volume, surface, and contour, which provides a predictable segmentation of a scene into its parts.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-12. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.