Label errors can significantly degrade model performance, making effective mechanisms crucial. Active error correction (AEC) addresses this by prioritizing data points for human re-labeling where corrections are expected to have significant impact. We extend AEC to distributed collaborative learning, where clients hold local data and a central server allocates labeling resources. Existing AEC methods assume centralized access and do not generalize to distributed settings. To overcome this, we use neural network weight gradients from client updates as proxies for local data and apply a Gaussian process in gradient space to strategically select clients for correction. Our method identifies gradient inconsistencies and encourages diversity through a computationally efficient rank-one Cholesky update. Experiments on eight benchmark datasets demonstrate the effectiveness of our approach.

Applying language models (LMs) to tables is challenging due to the mismatch between the two-dimensional structure of tables and the one-dimensional inputs expected by LMs.This mismatch forces linearization, making LMs particularly sensitive to irrelevant cells.Subtable selection mitigates this challenge by isolating question-relevant content prior to answer generation.However, existing approaches either rely on independent row or column selection, failing to capture cross-row and cross-column dependencies, or attempt global reasoning and face challenges similar to holistic table QA under noisy contexts.We propose PieTa (Piece of Table), a divide-and-conquer subtable selection framework that progressively aggregates locally selected evidence without requiring explicit global reasoning.PieTa uses an iterative, window-based multi-resolution process to construct compact subtables that capture global dependencies while limiting LM exposure to irrelevant content.Extensive experiments demonstrate that PieTa consistently outperforms prior subtable-based and holistic table QA approaches.Our code is available at https://github.com/MLV-group/ Piece-of-Table.git.

In privacy-preserving distributed learning environments, data stored on local clients cannot be shared with other clients or servers. We consider a new active learning problem setup for these environments, where the server aims to build a centralized model by distributing labeling budgets across different clients. Our algorithm identifies which clients and their data points warrant annotation by estimating the global impact of the resulting labels. We evaluate this impact by embedding the clients into the manifold of learner parameters, formed by the task learner's predictions on unlabeled data, and diffusing the reduction in predictive uncertainties caused by labeling. The algorithm effectively selects clients with high estimated impact while achieving diversity in client selection, all without accessing local client data. In experiments, our approach demonstrates substantial advancements when compared to adaptations of existing active learning algorithms.

We study multi-dataset training (MDT) for pose estimation, where skeletal heterogeneity presents a unique challenge that existing methods have yet to address. In traditional domains, e.g. regression and classification, MDT typically relies on dataset merging or multi-head supervision. However, the diversity of skeleton types and limited cross-dataset supervision complicate integration in pose estimation. To address these challenges, we introduce PoseBH, a new MDT framework that tackles keypoint heterogeneity and limited supervision through two key techniques. First, we propose nonparametric keypoint prototypes that learn within a unified embedding space, enabling seamless integration across skeleton types. Second, we develop a cross-type self-supervision mechanism that aligns keypoint predictions with keypoint embedding prototypes, providing supervision without relying on teacher-student models or additional augmentations. PoseBH substantially improves generalization across whole-body and animal pose datasets, including COCO-WholeBody, AP-10K, and APT-36K, while preserving performance on standard human pose benchmarks (COCO, MPII, and AIC). Furthermore, our learned key-point embeddings transfer effectively to hand shape estimation (InterHand2.6M) and human body shape estimation (3DPW). The code for PoseBH is available at: https://github.com/uyoung-jeong/PoseBH.

Recent rehearsal-free continual learning (CL) methods guided by prompts achieve strong performance on vision tasks with non-stationary data but remain resource-intensive, hindering real-world deployment. We introduce resource-efficient prompting (REP), which improves the computational and memory efficiency of prompt-based rehearsal-free methods while minimizing accuracy trade-offs. Our approach employs swift prompt selection to refine input data using a carefully provisioned model and introduces adaptive token merging AToM and adaptive layer dropping ALD for efficient prompt updates. AToM and ALD selectively skip data and model layers while preserving task-specific features during the learning of new tasks. Extensive experiments on multiple image classification datasets demonstrate REP’s superior resource efficiency over state-of-the-art rehearsal-free CL methods.

Re-training a deep learning model each time a single data point receives a new label is impractical due to the inherent complexity of the training process. Consequently, existing active learning (AL) algorithms tend to adopt a batch-based approach where, during each AL iteration, a set of data points is collectively chosen for annotation. However, this strategy frequently leads to redundant sampling, ultimately eroding the efficacy of the labeling procedure. In this paper, we introduce a new AL algorithm that harnesses the power of a Gaussian process surrogate in conjunction with the neural network principal learner. Our proposed model adeptly updates the surrogate learner for every new data instance, enabling it to emulate and capitalize on the continuous learning dynamics of the neural network without necessitating a complete retraining of the principal model for each individual label. Experiments on four benchmark datasets demonstrate that this approach yields significant enhancements, either rivaling or aligning with the performance of state-of-the-art techniques.

We study the distributed gradient aggregation problem where individual clients contribute to learning a central model by sharing parameter gradients constructed from local losses. However, errors in some gradients, caused by low-quality data or adversaries, can degrade the learning process when naively combined. Existing robust gradient aggregation approaches assume that local data represent the global data-generating distribution, which may not always apply to heterogeneous (non-i.i.d.) client data. We propose a new algorithm that can robustly aggregate gradients from potentially heterogeneous clients. Our approach leverages the manifold structure inherent in heterogeneous client gradients and evaluates gradient anomaly degrees by projecting them onto this manifold. This algorithm is implemented as a simple and efficient method that accumulates random projections within the subspace defined by the nearest neighbors within a gradient cloud. Our experiments demonstrate consistent performance improvements over state-of-the-art robust aggregation algorithms.

Neural field is an emerging paradigm in data representation that trains a neural network to approximate the given signal. A key obstacle that prevents its widespread adoption is the encoding speed-generating neural fields requires an overfitting of a neural network, which can take a significant number of SGD steps to reach the desired fidelity level. In this paper, we delve into the impacts of data transformations on the speed of neural field training, specifically focusing on how permuting pixel locations affect the convergence speed of SGD. Counterintuitively, we find that randomly permuting the pixel locations can considerably accelerate the training. To explain this phenomenon, we examine the neural field training through the lens of PSNR curves, loss landscapes, and error patterns. Our analyses suggest that the random pixel permutations remove the easy-to-fit patterns, which facilitate easy optimization in the early stage but hinder capturing fine details of the signal.11code: https://github.com/effl-lab/DT4Neural-Field

While there exists a great deal of labeled in-the-wild human data, the same is not true for animals. Manually creating new labels for the full range of animal species would take years of effort from the community. We are also now seeing the emerging potential for computer vision methods in areas like animal conservation, which is an additional motivation for this direction of research. Key to our approach is the ability to easily generate as many labeled training images as we desire across a range of different modalities. To achieve this, we present a new large scale canine motion capture dataset and parametric canine body and texture model. These are used to produce the first large scale, multi-domain, multi-task dataset for canine body analysis comprising of detailed synthetic labels on both real images and fully synthetic images in a range of realistic poses. We also introduce the first pose prior for animals in the form of a variational pose prior for canines which is used to fit the parametric model to images of canines. We demonstrate the effectiveness of our labels for training computer vision models on tasks such as parts-based segmentation and pose estimation and show such models can generalise to other animal species without additional training.

We consider distributed (gradient descent-based) learning scenarios where the server combines the gradients of learning objectives gathered from local clients. As individual data collection and learning environments can vary, some clients could transfer erroneous gradients e.g. due to ad-versarial data or gradient perturbations. Further, for data privacy and security, the identities of such affected clients are often unknown to the server. In such cases, naively ag-gregating the resulting gradients can mislead the learning process. We propose a new server-side learning algorithm that robustly combines gradients. Our algorithm embeds the local gradients into the manifold of normalized gradients and refines their combinations via simulating a diffusion process therein. The resulting algorithm is instantiated as a compu-tationally simple and efficient weighted gradient averaging algorithm. In the experiments with five classification and three regression benchmark datasets, our algorithm demon-strated significant performance improvements over existing robust gradient combination algorithms as well as the base-line uniform gradient averaging algorithm.

Estimating the pose and shape of hands and objects under interaction finds numerous applications including aug-mented and virtual reality. Existing approaches for hand and object reconstruction require explicitly defined physical constraints and known objects, which limits its application domains. Our algorithm is agnostic to object models, and it learns the physical rules governing hand-object interaction. This requires automatically inferring the shapes and physi-cal interaction of hands and (potentially unknown) objects. We seek to approach this challenging problem by proposing a collaborative learning strategy where two-branches of deep networks are learning from each other. Specifically, we transfer hand mesh information to the object branch and vice versa for the hand branch. The resulting optimi-sation (training) problem can be unstable, and we address this via two strategies: (i) attention-guided graph convo-lution which helps identify and focus on mutual occlusion and (ii) unsupervised associative loss which facilitates the transfer of information between the branches. Experiments using four widely-used benchmarks show that our frame-work achieves beyond state-of-the-art accuracy in 3D pose estimation, as well as recovers dense 3D hand and object shapes. Each technical component above contributes meaningfully in the ablation study.

Often, labeling large amount of data is challenging due to high labeling cost limiting the application domain of deep learning techniques. Active learning (AL) tackles this by querying the most informative samples to be annotated among unlabeled pool. Two promising directions for AL that have been recently explored are task-agnostic approach to select data points that are far from the current labeled pool and task-aware approach that relies on the perspective of task model. Unfortunately, the former does not exploit structures from tasks and the latter does not seem to well-utilize overall data distribution. Here, we propose task-aware variational adversarial AL (TA-VAAL) that modifies task-agnostic VAAL, that considered data distribution of both label and unlabeled pools, by relaxing task learning loss prediction to ranking loss prediction and by using ranking conditional generative adversarial network to embed normalized ranking loss information on VAAL. Our proposed TA-VAAL outperforms state-of-the-arts on various benchmark datasets for classifications with balanced / imbalanced labels as well as semantic segmentation and its task-aware and task-agnostic AL properties were confirmed with our in-depth analyses.

Three dimensional hand pose estimation has reached a level of maturity, enabling real-world applications for single-hand cases. However, accurate estimation of the pose of two closely interacting hands still remains a challenge as in this case, one hand often occludes the other. We present a new algorithm that accurately estimates hand poses in such a challenging scenario. The crux of our algorithm lies in a framework that jointly trains the estimators of interacting hands, leveraging their inter-dependence. Further, we employ a GAN-type discriminator of interacting hand pose that helps avoid physically implausible configurations, e.g. intersecting fingers, and exploit the visibility of joints to improve intermediate 2D pose estimation. We incorporate them into a single model that learns to detect hands and estimate their pose based on a unified criterion of pose estimation accuracy. To our knowledge, this is the first attempt to build an end-to-end network that detects and estimates the pose of two closely interacting hands (as well as single hands). In the experiments with three datasets representing challenging real-world scenarios, our algorithm demonstrated significant and consistent performance improvements over state-of-the-arts.

Given an imperfect predictor, we exploit additional features at test time to improve the predictions made, without retraining and without knowledge of the prediction function. This scenario arises if training labels or data are proprietary, restricted, or no longer available, or if training itself is prohibitively expensive. We assume that the additional features are useful if they exhibit strong statistical dependence to the underlying perfect predictor. Then, we empirically estimate and strengthen the statistical dependence between the initial noisy predictor and the additional features via manifold de-noising. As an example, we show that this approach leads to improvement in real-world visual attribute ranking.

Predictor combination is the problem of improving a task predictor using predictors of other tasks when the forms of individual predictors are unknown. Previous work approached this problem by nonparametrically assessing predictor relationships based on their joint evaluations on a shared sample. This limits their application to cases where all predictors are defined on the same task category, e.g. all predictors estimate attributes of shoes. We present a new predictor combination algorithm that overcomes this limitation. Our algorithm aligns the heterogeneous domains of different predictors in a shared latent space to facilitate comparisons of predictors independently of the domains on which they are originally defined. We facilitate this by a new data alignment scheme that matches data distributions across task categories. Based on visual attribute ranking experiments on datasets that span diverse task categories (e.g. shoes and animals), we demonstrate that our approach often significantly improves the performances of the initial predictors.

Despite recent successes in hand pose estimation, there yet remain challenges on RGB-based 3D hand pose estimation (HPE) under hand-object interaction (HOI) scenarios where severe occlusions and cluttered backgrounds exhibit. Recent RGB HOI benchmarks have been collected either in real or synthetic domain, however, the size of datasets is far from enough to deal with diverse objects combined with hand poses, and 3D pose annotations of real samples are lacking, especially for occluded cases. In this work, we propose a novel end-to-end trainable pipeline that adapts the hand-object domain to the single hand-only domain, while learning for HPE. The domain adaption occurs in image space via 2D pixel-level guidance by Generative Adversarial Network (GAN) and 3D mesh guidance by mesh renderer (MR). Via the domain adaption in image space, not only 3D HPE accuracy is improved, but also HOI input images are translated to segmented and de-occluded hand-only images. The proposed method takes advantages of both the guidances: GAN accurately aligns hands, while MR effectively fills in occluded pixels. The experiments using Dexter-Object, Ego-Dexter and HO3D datasets show that our method significantly outperforms state-of-the-arts trained by hand-only data and is comparable to those supervised by HOI data. Note our method is trained primarily by hand-only images with pose labels, and HOI images without pose labels.

The automatic extraction of animal 3D pose from images without markers is of interest in a range of scientific fields. Most work to date predicts animal pose from RGB images, based on 2D labelling of joint positions. However, due to the difficult nature of obtaining training data, no ground truth dataset of 3D animal motion is available to quantitatively evaluate these approaches. In addition, a lack of 3D animal pose data also makes it difficult to train 3D pose-prediction methods in a similar manner to the popular field of body-pose prediction. In our work, we focus on the problem of 3D canine pose estimation from RGBD images, recording a diverse range of dog breeds with several Microsoft Kinect v2s, simultaneously obtaining the 3D ground truth skeleton via a motion capture system. We generate a dataset of synthetic RGBD images from this data. A stacked hourglass network is trained to predict 3D joint locations, which is then constrained using prior models of shape and pose. We evaluate our model on both synthetic and real RGBD images and compare our results to previously published work fitting canine models to images. Finally, despite our training set consisting only of dog data, visual inspection implies that our network can produce good predictions for images of other quadrupeds - e.g. horses or cats - when their pose is similar to that contained in our training set.

Estimating 3D hand meshes from single RGB images is challenging, due to intrinsic 2D-3D mapping ambiguities and limited training data. We adopt a compact parametric 3D hand model that represents deformable and articulated hand meshes. To achieve the model fitting to RGB images, we investigate and contribute in three ways: 1) Neural rendering: inspired by recent work on human body, our hand mesh estimator (HME) is implemented by a neural network and a differentiable renderer, supervised by 2D segmentation masks and 3D skeletons. HME demonstrates good performance for estimating diverse hand shapes and improves pose estimation accuracies. 2) Iterative testing refinement: Our fitting function is differentiable. We iteratively refine the initial estimate using the gradients, in the spirit of iterative model fitting methods like ICP. The idea is supported by the latest research on human body. 3) Self-data augmentation: collecting sized RGB-mesh (or segmentation mask)-skeleton triplets for training is a big hurdle. Once the model is successfully fitted to input RGB images, its meshes i.e. shapes and articulations, are realistic, and we augment view-points on top of estimated dense hand poses. Experiments using three RGB-based benchmarks show that our framework offers beyond state-of-the-art accuracy in 3D pose estimation, as well as recovers dense 3D hand shapes. Each technical component above meaningfully improves the accuracy in the ablation study.

We present a new predictor combination algorithm that improves a given task predictor based on potentially relevant reference predictors. Existing approaches are limited in that, to discover the underlying task dependence, they either require known parametric forms of all predictors or access to a single fixed dataset on which all predictors are jointly evaluated. To overcome these limitations, we design a new non-parametric task dependence estimation procedure that automatically aligns evaluations of heterogeneous predictors across disjoint feature sets. Our algorithm is instantiated as a robust manifold diffusion process that jointly refines the estimated predictor alignments and the corresponding task dependence. We apply this algorithm to the relative attributes ranking problem and demonstrate that it not only broadens the application range of predictor combination approaches but also outperforms existing methods even when applied to classical predictor combination settings.

We investigate filter level sparsity that emerges in convolutional neural networks (CNNs) which employ Batch Normalization and ReLU activation, and are trained with adaptive gradient descent techniques and L2 regularization or weight decay. We conduct an extensive experimental study casting our initial findings into hypotheses and conclusions about the mechanisms underlying the emergent filter level sparsity. This study allows new insight into the performance gap obeserved between adapative and non-adaptive gradient descent methods in practice. Further, analysis of the effect of training strategies and hyperparameters on the sparsity leads to practical suggestions in designing CNN training strategies enabling us to explore the tradeoffs between feature selectivity, network capacity, and generalization performance. Lastly, we show that the implicit sparsity can be harnessed for neural network speedup at par or better than explicit sparsification / pruning approaches, with no modifications to the typical training pipeline required.

Crucial to the success of training a depth-based 3D hand pose estimator (HPE) is the availability of comprehensive datasets covering diverse camera perspectives, shapes, and pose variations. However, collecting such annotated datasets is challenging. We propose to complete existing databases by generating new database entries. The key idea is to synthesize data in the skeleton space (instead of doing so in the depth-map space) which enables an easy and intuitive way of manipulating data entries. Since the skeleton entries generated in this way do not have the corresponding depth map entries, we exploit them by training a separate hand pose generator (HPG) which synthesizes the depth map from the skeleton entries. By training the HPG and HPE in a single unified optimization framework enforcing that 1) the HPE agrees with the paired depth and skeleton entries; and 2) the HPG-HPE combination satisfies the cyclic consistency (both the input and the output of HPG-HPE are skeletons) observed via the newly generated unpaired skeletons, our algorithm constructs a HPE which is robust to variations that go beyond the coverage of the existing database. Our training algorithm adopts the generative adversarial networks (GAN) training process. As a by-product, we obtain a hand pose discriminator (HPD) that is capable of picking out realistic hand poses. Our algorithm exploits this capability to refine the initial skeleton estimates in testing, further improving the accuracy. We test our algorithm on four challenging benchmark datasets (ICVL, MSRA, NYU and Big Hand 2.2M datasets) and demonstrate that our approach outperforms or is on par with state-of-the-art methods quantitatively and qualitatively.

We present a new multi-task learning (MTL) approach that can be applied to multiple heterogeneous task estimators. Our motivation is that the best task estimator could change depending on the task itself. For example, we may have a deep neural network for the first task and a Gaussian process for the second task. Classical MTL approaches cannot handle this case, as they require the same model or even the same parameter types for all tasks. We tackle this by considering task-specific estimators as random variables. Then, the task relationships are discovered by measuring the statistical dependence between each pair of random variables. By doing so, our model is independent of the parametric nature of each task, and is even agnostic to the existence of such parametric formulation. We compare our algorithm with existing MTL approaches on challenging real world ranking and regression datasets, and show that our approach achieves comparable or better performance without knowing the parametric form.

Current unsupervised image-to-image translation techniques struggle to focus their attention on individual objects without altering the background or the way multiple objects interact within a scene. Motivated by the important role of attention in human perception, we tackle this limitation by introducing unsupervised attention mechanisms which are jointly adversarially trained with the generators and discriminators. We empirically demonstrate that our approach is able to attend to relevant regions in the image without requiring any additional supervision, and that by doing so it achieves more realistic mappings compared to recent approaches.

We present an algorithm for test-time combination of a set of reference predictors with unknown parametric forms. Existing multi-task and transfer learning algorithms focus on training-time transfer and combination, where the parametric forms of predictors are known and shared. However, when the parametric form of a predictor is unknown, e.g., for a human predictor or a predictor in a precompiled library, existing algorithms are not applicable. Instead, we empirically evaluate predictors on sampled data points to measure distances between different predictors. This embeds the set of reference predictors into a Riemannian manifold, upon which we perform manifold denoising to obtain the refined predictor. This allows our approach to make no assumptions about the underlying predictor forms. Our test-time combination algorithm equals or outperforms existing multi-task and transfer learning algorithms on challenging real-world datasets, without introducing specific model assumptions.

Online action detection (OAD) is challenging since 1) robust yet computationally expensive features cannot be straightforwardly used due to the real-time processing requirements and 2) the localization and classification of actions have to be performed even before they are fully observed. We propose a new random forest (RF)-based online action detection framework that addresses these challenges. Our algorithm uses computationally efficient skeletal joint features. High accuracy is achieved by using robust convolutional neural network (CNN)-based features which are extracted from the raw RGBD images, plus the temporal relationships between the current frame of interest, and the past and futures frames. While these high-quality features are not available in real-time testing scenario, we demonstrate that they can be effectively exploited in training RF classifiers: We use these spatio-temporal contexts to craft RF's new split functions improving RFs' leaf node statistics. Experiments with challenging MSRAction3D, G3D, and OAD datasets demonstrate that our algorithm significantly improves the accuracy over the state-of-the-art on-line action detection algorithms while achieving the real-time efficiency of existing skeleton-based RF classifiers.

In many learning tasks, the structure of the target space of a function holds rich information about the relationships between evaluations of functions on different data points. Existing approaches attempt to exploit this relationship information implicitly by enforcing smoothness on function evaluations only. However, what happens if we explicitly regularize the relationships between function evaluations? Inspired by homophily, we regularize based on a smooth relationship function, either defined from the data or with labels. In experiments, we demonstrate that this significantly improves the performance of state-of-the-art algorithms in semi-supervised classification and in spectral data embedding for constrained clustering and dimensionality reduction.

The common graph Laplacian regularizer is well-established in semi-supervised learning and spectral dimensionality reduction. However, as a first-order regularizer, it can lead to degenerate functions in high-dimensional manifolds. The iterated graph Laplacian enables high-order regularization, but it has a high computational complexity and so cannot be applied to large problems. We introduce a new regularizer which is globally high order and so does not suffer from the degeneracy of the graph Laplacian regularizer, but is also sparse for efficient computation in semi-supervised learning applications. We reduce computational complexity by building a local first-order approximation of the manifold as a surrogate geometry, and construct our high-order regularizer based on local derivative evaluations therein. Experiments on human body shape and pose analysis demonstrate the effectiveness and efficiency of our method.

Existing approaches for diffusion on graphs, e.g., for label propagation, are mainly focused on isotropic diffusion, which is induced by the commonly-used graph Laplacian regularizer. Inspired by the success of diffusivity tensors for anisotropic diffusion in image processing, we presents anisotropic diffusion on graphs and the corresponding label propagation algorithm. We develop positive definite diffusivity operators on the vector bundles of Riemannian manifolds, and discretize them to diffusivity operators on graphs. This enables us to easily define new robust diffusivity operators which significantly improve semi-supervised learning performance over existing diffusion algorithms.

Abstract We present a method for capturing the skeletal motions of humans using a sparse set of potentially moving cameras in an uncontrolled environment. Our approach is able to track multiple people even in front of cluttered and non‐static backgrounds, and unsynchronized cameras with varying image quality and frame rate. We completely rely on optical information and do not make use of additional sensor information (e.g. depth images or inertial sensors). Our algorithm simultaneously reconstructs the skeletal pose parameters of multiple performers and the motion of each camera. This is facilitated by a new energy functional that captures the alignment of the model and the camera positions with the input videos in an analytic way. The approach can be adopted in many practical applications to replace the complex and expensive motion capture studios with few consumer‐grade cameras even in uncontrolled outdoor scenes. We demonstrate this based on challenging multi‐view video sequences that are captured with unsynchronized and moving (e.g. mobile‐phone or GoPro) cameras.

Improving the quality of degraded images is a key problem in image processing, but the breadth of the problem leads to domain-specific approaches for tasks such as super-resolution and compression artifact removal. Recent approaches have shown that a general approach is possible by learning application-specific models from examples; however, learning models sophisticated enough to generate high-quality images is computationally expensive, and so specific per-application or per-dataset models are impractical. To solve this problem, we present an efficient semi-local approximation scheme to large-scale Gaussian processes. This allows efficient learning of task-specific image enhancements from example images without reducing quality. As such, our algorithm can be easily customized to specific applications and datasets, and we show the efficiency and effectiveness of our approach across five domains: single-image super-resolution for scene, human face, and text images, and artifact removal in JPEG- and JPEG 2000-encoded images.

Motion-tracked real-time character control is important for games and VR, but current solutions are limited: retargeting is hard for non-human characters, with locomotion bound to the sensing volume; and pose mappings are ambiguous with difficult dynamic motion control. We robustly estimate wave properties --- amplitude, frequency, and phase ---for a set of interactively-defined gestures by mapping user motions to a low-dimensional independent representation. The mapping separates simultaneous or intersecting gestures, and extrapolates gesture variations from single training examples. For animations such as locomotion, wave properties map naturally to stride length, step frequency, and progression, and allow smooth transitions from standing, to walking, to running. Interpolating out-of-phase locomotions is hard, e.g., quadruped legs between walks and runs switch phase, so we introduce a new time-interpolation scheme to reduce artifacts. These improvements to real-time motion-tracked character control are important for common cyclic animations. We validate this in a user study, and show versatility to apply to part- and full-body motions across a variety of sensors.

Abstract It is now possible to capture the 3D motion of the human body on consumer hardware and to puppet in real time skeleton‐based virtual characters. However, many characters do not have humanoid skeletons. Characters such as spiders and caterpillars do not have boned skeletons at all, and these characters have very different shapes and motions. In general, character control under arbitrary shape and motion transformations is unsolved ‐ how might these motions be mapped? We control characters with a method which avoids the rigging‐skinning pipeline — source and target characters do not have skeletons or rigs. We use interactively‐defined sparse pose correspondences to learn a mapping between arbitrary 3D point source sequences and mesh target sequences. Then, we puppet the target character in real time. We demonstrate the versatility of our method through results on diverse virtual characters with different input motion controllers. Our method provides a fast, flexible, and intuitive interface for arbitrary motion mapping which provides new ways to control characters for real‐time animation.

One fundamental assumption in object recognition as well as in other computer vision and pattern recognition problems is that the data generation process lies on a manifold and that it respects the intrinsic geometry of the manifold. This assumption is held in several successful algorithms for diffusion and regularization, in particular, in graph-Laplacian-based algorithms. We claim that the performance of existing algorithms can be improved if we additionally account for how the manifold is embedded within the ambient space, i.e., if we consider the extrinsic geometry of the manifold. We present a procedure for characterizing the extrinsic (as well as intrinsic) curvature of a manifold M which is described by a sampled point cloud in a high-dimensional Euclidean space. Once estimated, we use this characterization in general diffusion and regularization on M, and form a new regularizer on a point cloud. The resulting re-weighted graph Laplacian demonstrates superior performance over classical graph Laplacian in semi-supervised learning and spectral clustering.

Emerging interfaces for video collections of places attempt to link similar content with seamless transitions. However, the automatic computer vision techniques that enable these transitions have many failure cases which lead to artifacts in the final rendered transition. Under these conditions, which transitions are preferred by participants and which artifacts are most objectionable? We perform an experiment with participants comparing seven transition types, from movie cuts and dissolves to image-based warps and virtual camera transitions, across five scenes in a city. This document describes how we condition this experiment on slight and considerable view change cases, and how we analyze the feedback from participants to find their preference for transition types and artifacts. We discover that transition preference varies with view change, that automatic rendered transitions are significantly preferred even with some artifacts, and that dissolve transitions are comparable to less-sophisticated rendered transitions. This leads to insights into what visual features are important to maintain in a rendered transition, and to an artifact ordering within our transitions.

High dynamic range reconstruction of dynamic scenes requires careful handling of dynamic objects to prevent ghosting. However, in a recent review, Srikantha et al. [2012] conclude that "there is no single best method and the selection of an approach depends on the user's goal". We attempt to solve this problem with a novel approach that models the noise distribution of color values. We estimate the likelihood that a pair of colors in different images are observations of the same irradiance, and we use a Markov random field prior to reconstruct irradiance from pixels that are likely to correspond to the same static scene object. Dynamic content is handled by selecting a single low dynamic range source image and hand-held capture is supported through homography-based image alignment. Our noise-based reconstruction method achieves better ghost detection and removal than state-of-the-art methods for cluttered scenes with large object displacements. As such, our method is broadly applicable and helps move the field towards a single method for dynamic scene HDR reconstruction.

In this paper, we describe a framework for learning-based image enhancement. At the core of our algorithm lies a generic regularization framework that comprises a prior on natural images, as well as an application-specific conditional model based on Gaussian processes. In contrast to prior learning-based approaches, our algorithm can instantly learn task-specific degradation models from sample images which enables users to easily adopt the algorithm to a specific problem and data set of interest. This is facilitated by our efficient approximation scheme of large-scale Gaussian processes. We demonstrate the efficiency and effectiveness of our approach by applying it to two example enhancement applications: single-image super-resolution as well as artifact removal in JPEG-encoded images.

We present a new spatio-temporal method for markerless motion capture. We reconstruct the pose and motion of a character from a multi-view video sequence without requiring the cameras to be synchronized and without aligning captured frames in time. By formulating the model-to-image similarity measure as a temporally continuous functional, we are also able to reconstruct motion in much higher temporal detail than was possible with previous synchronized approaches. By purposefully running cameras unsynchronized we can capture even very fast motion at speeds that off-the-shelf but high quality cameras provide.

Abstract Removing dynamic objects from videos is an extremely challenging problem that even visual effects professionals often solve with time‐consuming manual frame‐by‐frame editing. We propose a new approach to video completion that can deal with complex scenes containing dynamic background and non‐periodical moving objects. We build upon the idea that the spatio‐temporal hole left by a removed object can be filled with data available on other regions of the video where the occluded objects were visible. Video completion is performed by solving a large combinatorial problem that searches for an optimal pattern of pixel offsets from occluded to unoccluded regions. Our contribution includes an energy functional that generalizes well over different scenes with stable parameters, and that has the desirable convergence properties for a graph‐cut‐based optimization. We provide an interface to guide the completion process that both reduces computation time and allows for efficient correction of small errors in the result. We demonstrate that our approach can effectively complete complex, high‐resolution occlusions that are greater in difficulty than what existing methods have shown.

Abstract Synthetically generating images and video frames of complex 3D scenes using some photo‐realistic rendering software is often prone to artifacts and requires expert knowledge to tune the parameters. The manual work required for detecting and preventing artifacts can be automated through objective quality evaluation of synthetic images. Most practical objective quality assessment methods of natural images rely on a ground‐truth reference, which is often not available in rendering applications. While general purpose no‐reference image quality assessment is a difficult problem, we show in a subjective study that the performance of a dedicated no‐reference metric as presented in this paper can match the state‐of‐the‐art metrics that do require a reference. This level of predictive power is achieved exploiting information about the underlying synthetic scene (e.g., 3D surfaces, textures) instead of merely considering color, and training our learning framework with typical rendering artifacts. We show that our method successfully detects various non‐trivial types of artifacts such as noise and clamping bias due to insufficient virtual point light sources, and shadow map discretization artifacts. We also briefly discuss an inpainting method for automatic correction of detected artifacts.

The abundance of mobile devices and digital cameras with video capture makes it easy to obtain large collections of video clips that contain the same location, environment, or event. However, such an unstructured collection is difficult to comprehend and explore. We propose a system that analyzes collections of unstructured but related video data to create a Videoscape : a data structure that enables interactive exploration of video collections by visually navigating -- spatially and/or temporally -- between different clips. We automatically identify transition opportunities, or portals . From these portals, we construct the Videoscape, a graph whose edges are video clips and whose nodes are portals between clips. Now structured, the videos can be interactively explored by walking the graph or by geographic map. Given this system, we gauge preference for different video transition styles in a user study, and generate heuristics that automatically choose an appropriate transition style. We evaluate our system using three further user studies, which allows us to conclude that Videoscapes provides significant benefits over related methods. Our system leads to previously unseen ways of interactive spatio-temporal exploration of casually captured videos, and we demonstrate this on several video collections.

This paper proposes a framework for single-image super-resolution. The underlying idea is to learn a map from input low-resolution images to target high-resolution images based on example pairs of input and output images. Kernel ridge regression (KRR) is adopted for this purpose. To reduce the time complexity of training and testing for KRR, a sparse solution is found by combining the ideas of kernel matching pursuit and gradient descent. As a regularized solution, KRR leads to a better generalization than simply storing the examples as has been done in existing example-based algorithms and results in much less noisy images. However, this may introduce blurring and ringing artifacts around major edges as sharp changes are penalized severely. A prior model of a generic image class which takes into account the discontinuity property of images is adopted to resolve this problem. Comparison with existing algorithms shows the effectiveness of the proposed method.

Semi-supervised regression based on the graph Laplacian suffers from the fact that the solution is biased towards a constant and the lack of extrapolating power. Outgoing from these observations we propose to use the second-order Hessian energy for semi-supervised regression which overcomes both of these problems, in particular, if the data lies on or close to a low-dimensional submanifold in the feature space, the Hessian energy prefers functions which vary ``linearly with respect to the natural parameters in the data. This property makes it also particularly suited for the task of semi-supervised dimensionality reduction where the goal is to find the natural parameters in the data based on a few labeled points. The experimental result suggest that our method is superior to semi-supervised regression using Laplacian regularization and standard supervised methods and is particularly suited for semi-supervised dimensionality reduction.

This paper presents a fully automated algorithm for reconstructing a textured 3D model of a face from a single photograph or a raw video stream. The algorithm is based on a combination of Support Vector Machines (SVMs) and a Morphable Model of 3D faces. After SVM face detection, individual facial features are detected using a novel regression- and classification-based approach, and probabilistically plausible configurations of features are selected to produce a list of candidates for several facial feature positions. In the next step, the configurations of feature points are evaluated using a novel criterion that is based on a Morphable Model and a combination of linear projections. To make the algorithm robust with respect to head orientation, this process is iterated while the estimate of pose is refined. Finally, the feature points initialize a model-fitting procedure of the Morphable Model. The result is a highresolution 3D surface model.

Most existing sparse Gaussian process (g.p.) models seek computational advantages by basing their computations on a set of m basis functions that are the covariance function of the g.p. with one of its two inputs fixed. We generalise this for the case of Gaussian covariance function, by basing our computations on m Gaussian basis functions with arbitrary diagonal covariance matrices (or length scales). For a fixed number of basis functions and any given criteria, this additional flexibility permits approximations no worse and typically better than was previously possible. We perform gradient based optimisation of the marginal likelihood, which costs O(m2n) time where n is the number of data points, and compare the method to various other sparse g.p. methods. Although we focus on g.p. regression, the central idea is applicable to all kernel based algorithms, and we also provide some results for the support vector machine (s.v.m.) and kernel ridge regression (k.r.r.). Our approach outperforms the other methods, particularly for the case of very few basis functions, i. e. a very high sparsity ratio.

In recent years, Kernel Principal Component Analysis (KPCA) has been suggested for various image processing tasks requiring an image model such as, e.g., denoising or compression. The original form of KPCA, however, can be only applied to strongly restricted image classes due to the limited number of training examples that can be processed. We therefore propose a new iterative method for performing KPCA, the Kernel Hebbian Algorithm which iteratively estimates the Kernel Principal Components with only linear order memory complexity. In our experiments, we compute models for complex image classes such as faces and natural images which require a large number of training examples. The resulting image models are tested in single-frame super-resolution and denoising applications. The KPCA model is not specifically tailored to these tasks; in fact, the same model can be used in super-resolution with variable input resolution, or denoising with unknown noise characteristics. In spite of this, both super-resolution and denoising performance are comparable to existing methods.

The current paper presents a novel texture-based method for detecting texts in images. A support vector machine (SVM) is used to analyze the textural properties of texts. No external texture feature extraction module is used, but rather the intensities of the raw pixels that make up the textural pattern are fed directly to the SVM, which works well even in high-dimensional spaces. Next, text regions are identified by applying a continuously adaptive mean shift algorithm (CAMSHIFT) to the results of the texture analysis. The combination of CAMSHIFT and SVMs produces both robust and efficient text detection, as time-consuming texture analyses for less relevant pixels are restricted, leaving only a small part of the input image to be texture-analyzed.

This paper presents a real-time face recognition system. For the system to be real time, no external time-consuming feature extraction method is used, rather the gray-level values of the raw pixels that make up the face pattern are fed directly to the recognizer. In order to absorb the resulting high dimensionality of the input space, support vector machines (SVMs), which are known to work well even in high-dimensional space, are used as the face recognizer. Furthermore, a modified form of polynomial kernel (local correlation kernel) is utilized to take account of prior knowledge about facial structures and is used as the alternative feature extractor. Since SVMs were originally developed for two-class classification, their basic scheme is extended for multiface recognition by adopting one-per-class decomposition. In order to make a final classification from several one-per-class SVM outputs, a neural network (NN) is used as the arbitrator. Experiments with ORL database show a recognition rate of 97.9% and speed of 0.22 seconds per face with 40 classes.

This paper investigates the application of support vector machines (SVMs) in texture classification. Instead of relying on an external feature extractor, the SVM receives the gray-level values of the raw pixels, as SVMs can generalize well even in high-dimensional spaces. Furthermore, it is shown that SVMs can incorporate conventional texture feature extraction methods within their own architecture, while also providing solutions to problems inherent in these methods. One-against-others decomposition is adopted to apply binary SVMs to multitexture classification, plus a neural network is used as an arbitrator to make final classifications from several one-against-others SVM outputs. Experimental results demonstrate the effectiveness of SVMs in texture classification.

A kernel principal component analysis (PCA) was previously proposed as a nonlinear extension of a PCA. The basic idea is to first map the input space into a feature space via nonlinear mapping and then compute the principal components in that feature space. This article adopts the kernel PCA as a mechanism for extracting facial features. Through adopting a polynomial kernel, the principal components can be computed within the space spanned by high-order correlations of input pixels making up a facial image, thereby producing a good performance.