Showing papers in "Computer Vision and Image Understanding in 2013"

TL;DR: The survey introduced in this paper tries to cover the lack of a complete description of the most important public datasets for video-based human activity and action recognition and to guide researchers in the election of themost suitable dataset for benchmarking their algorithms.

TL;DR: Gradient Field HOG is described; an adapted form of the HOG descriptor suitable for Sketch Based Image Retrieval (SBIR) and incorporated into a Bag of Visual Words retrieval framework, and shown to consistently outperform retrieval versus SIFT, multi-resolution HOG, Self Similarity, Shape Context and Structure Tensor.

TL;DR: It is argued that the next step in the evolution of object recognition algorithms will require radical and bold steps forward in terms of the object representations, as well as the learning and inference algorithms used.

TL;DR: The capabilities of SDALF in encoding peculiar aspects of an individual are shown, focusing on its robustness properties across dramatic low resolution images, in presence of occlusions and pose changes, and variations of viewpoints and scene illumination.

TL;DR: This survey provides a compact and informative summary of the major literature in this research topic, which substantially enhances the expressiveness of graph-based models and expands the domain of solvable problems.

TL;DR: B BossaNova is proposed, a novel representation for content-based concept detection in images and videos, which enriches the Bag-of-Words model, and is compact and simple to compute.

TL;DR: The proposed method outperformed all other approaches based on BOV that do not account for contextual information and imposes spatial and temporal constraints on the video volumes so that an inference mechanism can estimate the probability density functions of their arrangements.

TL;DR: This work presents a novel tracking-by-detection approach based on the generalized Hough-transform and demonstrates it for a variety of previously unknown objects even under heavy non-rigid transformations, partial occlusions, scale changes, and rotations.

TL;DR: The meshSIFT algorithm and its use for 3D face recognition is presented, and it is demonstrated that symmetrising the feature descriptors allows comparing two 3D facial surfaces with limited or no overlap.

TL;DR: A survey of shape modeling applications to cardiac image analysis from MRI, CT, echocardiography, PET, and SPECT is presented and aims to introduce new methodologies in this field, classify major contributions in image-based cardiac modeling, and provide a tutorial to beginners to initiate their own studies.

TL;DR: The present work proposes a novel probabilistic approach based on genetic algorithms to reconstruct iris images from binary templates and analyzes the similarity between the reconstructed synthetic iris image and the original one.

TL;DR: MSGVF is developed so that when the contour reaches equilibrium, the various forces resulting from the different energy terms are balanced and the smoothness constraint of image pixels is kept so that over- or under-segmentation can be reduced.

TL;DR: A new method based on learned bone-structure edge detectors and a coarse-to-fine deformable surface model is proposed to segment and identify vertebrae in 3D CT thoracic images and achieves a success rate comparable or slightly better than state-of-the-art.

TL;DR: A working process to help recognize tree species, starting from a picture of a leaf in a complex natural background, and high-level geometrical descriptors that prove to achieve better performance than more generic and statistical shape descriptors alone are developed.

TL;DR: An extensive evaluation study of different strategies for computing adaptive support weights in local stereo matching, including the original bilateral filter-based weights, as well as more recent approaches based on geodesic distances or on the guided filter.

TL;DR: A number of techniques for generating mid-level features, including two variants of Soft Assignment, Locality-constrained Linear Coding, and Sparse Coding are reviewed, and an intuitive concept of improved pooling is introduced.

TL;DR: This paper presents an adaptive spatial information-theoretic fuzzy clustering algorithm to improve the robustness of the conventional fuzzy c-means (FCM) clustering algorithms for image segmentation and it yields better segmentation results when compared to the conventional FCM approach.

TL;DR: A novel algorithm to detect text information from natural scene images that achieves the state-of-the-art performance on scene text classification and detection, and significantly outperforms the existing algorithms for character identification.

TL;DR: This work introduces a segmentation method based on a statistical shape model obtained with a principal component analysis (PCA) on a set of representative shapes of the RV that has been applied on 248 MR images of a publicly available dataset and shows that encouraging results can be obtained.

TL;DR: The robustness of MBD to noise and blur, as well as its stability with respect to the change of a position of points within the same object (or its background) are investigated, suggesting that the proposed minimum barrier distance is potentially useful in different imaging tasks, such as image segmentation.

TL;DR: A multiple object segmentation method using a novel and efficient object representation for both two and three dimensions is presented, which guarantees object relationships and topology, prevents overlaps and gaps, enables boundary-specific speeds, and has a computationally efficient evolution scheme that is largely independent of the number of objects.

TL;DR: A novel external force, called adaptive diffusion flow (ADF), with adaptive diffusion strategies according to the characteristics of an image region in the parametric active contour model framework for image segmentation is proposed.

TL;DR: A novel method that can rapidly detect an object's 3D rigid motion or deformation from a 2D projection image or a small set thereof, requires only a few seconds and yields good results in localizing a tumor under rigid motion in the head and neck and under respiratory deformation in the lung.

TL;DR: This work proposes two types of local descriptors based on Gaussian derivatives filters, both of them have the property of continuous rotation invariance and achieves remarkable performance to classify the rotated textures.

TL;DR: This paper presents an automatic LV myocardial boundary segmentation method using the parametric active contour model (or snake model), and a novel shape-similarity based energy is proposed to prevent the snake contour from being strapped in faulty edges and to preserve weak boundaries.

TL;DR: A highly efficient expectation-maximization (EM) algorithm is presented, based on minimum message length (MML) formulation, for the unsupervised learning of the proposed model's parameters, and its performance in two interesting applications namely pedestrian detection and multiple target tracking is studied.

TL;DR: A generic online multi-target track-before-detect (MT-TBD) that is applicable on confidence maps used as observations and a probabilistic model of target birth and death based on a Markov Random Field applied to the particle IDs is proposed.

TL;DR: A three-phase gait recognition method that analyses the spatio-temporal shape and dynamic motion characteristics of a human subject's silhouettes to identify the subject in the presence of most of the challenging factors that affect existing gait recognized systems is presented.

TL;DR: This work considers two dual adjunctions between the edge set and the vertex set of an arbitrary (unweighted) graph G to recover the classical notion of a dilation/erosion of a subset of the vertices of G and proposes several new openings, closings, granulometries and alternate filters.

TL;DR: This paper presents an approach to extract curvilinear structures (lines) and their widths from two-dimensional images with high accuracy and shows that very accurate results can be achieved on real data if the appropriate line model is used.