Histograms of Oriented Gradients (HoGs) provide excellent results in object detection and verification. However, their demanding processing requirements bound their applicability in some critical real-time scenarios, such as for video-based on-board vehicle detection systems. In this work, an efficient HOG configuration for pose-based on-board vehicle verification is proposed, which alleviates both the processing requirements and required feature vector length without reducing classification performance. The impact on classification of some critical configuration and processing parameters is in depth analyzed to propose a baseline efficient descriptor. Based on the analysis of its cells contribution to classification, new view-dependent cell-configuration patterns are proposed, resulting in reduced descriptors which provide an excellent balance between performance and computational requirements, rendering higher verification rates than other works in the literature.

Low-cost depth cameras, such as Microsoft Kinect, have completely changed the world of human-computer interaction through controller-free gaming applications. Depth data provided by the Kinect sensor presents several noise-related problems that have to be tackled to improve the accuracy of the depth data, thus obtaining more reliable game control platforms and broadening its applicability. In this paper, we present a depth-color fusion strategy for 3-D modeling of indoor scenes with Kinect. Accurate depth and color models of the background elements are iteratively built, and used to detect moving objects in the scene. Kinect depth data is processed with an innovative adaptive joint-bilateral filter that efficiently combines depth and color by analyzing an edge-uncertainty map and the detected foreground regions. Results show that the proposed approach efficiently tackles main Kinect data problems: distance-dependent depth maps, spatial noise, and temporal random fluctuations are dramatically reduced; objects depth boundaries are refined, and nonmeasured depth pixels are interpolated. Moreover, a robust depth and color background model and accurate moving objects silhouette are generated.

Vehicle detection based on image analysis has attracted increasing attention in recent years due to its low cost, flexibility, and potential toward collision avoidance. In particular, vehicle verification is especially challenging on account of the heterogeneity of vehicles in color, size, pose, etc. Image-based vehicle verification is usually addressed as a supervised classification problem. Specifically, descriptors using Gabor filters have been reported to show good performance in this task. However, Gabor functions have a number of drawbacks relating to their frequency response. The main contribution of this paper is the proposal and evaluation of a new descriptor based on the alternative family of log-Gabor functions for vehicle verification, as opposed to existing Gabor filter-based descriptors. These filters are theoretically superior to Gabor filters as they can better represent the frequency properties of natural images. As a second contribution, and in contrast to existing approaches, which transfer the standard configuration of filters used for other applications to the vehicle classification task, an in-depth analysis of the required filter configuration by both Gabor and log-Gabor descriptors for this particular application is performed for fair comparison. The extensive experiments conducted in this paper confirm that the proposed log-Gabor descriptor significantly outperforms the standard Gabor filter for image-based vehicle verification.

Video-based vehicle detection is the focus of increasing interest due to its potential towards collision avoidance. In particular, vehicle verification is especially challenging due to the enormous variability of vehicles in size, color, pose, etc. In this paper, a new approach based on supervised learning using Principal Component Analysis (PCA) is proposed that addresses the main limitations of existing methods. Namely, in contrast to classical approaches which train a single classifier regardless of the relative position of the candidate (thus ignoring valuable pose information), a region-dependent analysis is performed by considering four different areas. In addition, a study on the evolution of the classification performance according to the dimensionality of the principal subspace is carried out using PCA features within a SVM-based classification scheme. Indeed, the experiments performed on a publicly available database prove that PCA dimensionality requirements are region-dependent. Hence, in this work, the optimal configuration is adapted to each of them, rendering very good vehicle verification results.

In this paper, we present a depth-color scene modeling strategy for indoors 3D contents generation. It combines depth and visual information provided by a low-cost active depth camera to improve the accuracy of the acquired depth maps considering the different dynamic nature of the scene elements. Accurate depth and color models of the scene background are iteratively built, and used to detect moving elements in the scene. The acquired depth data is continuously processed with an innovative joint-bilateral filter that efficiently combines depth and visual information thanks to the analysis of an edge-uncertainty map and the detected foreground regions. The main advantages of the proposed approach are: removing depth maps spatial noise and temporal random fluctuations; refining depth data at object boundaries, generating iteratively a robust depth and color background model and an accurate moving object silhouette.

One of the main challenges for intelligent vehicles is the capability of detecting other vehicles in their environment, which constitute the main source of accidents. Specifically, many methods have been proposed in the literature for video-based vehicle detection. Most of them perform supervised classification using some appearance-related feature, in particular, symmetry has been extensively utilized. However, an in-depth analysis of the classification power of this feature is missing. As a first contribution of this paper, a thorough study of the classification performance of symmetry is presented within a Bayesian decision framework. This study reveals that the performance of symmetry-based classification is very limited. Therefore, as a second contribution, a new gradient-based descriptor is proposed for vehicle detection. This descriptor exploits the known rectangular structure of vehicle rears within a Histogram of Gradients (HOG)-based framework. Experiments show that the proposed descriptor outperforms largely symmetry as a feature for vehicle verification, achieving classification rates over 90%.

In the field of detection and monitoring of dynamic objects in quasi-static scenes, background subtraction techniques where background is modeled at pixel-level, although showing very significant limitations, are extensively used. In this work we propose a novel approach to background modeling that operates at region-level in a wavelet based multi-resolution framework. Based on a segmentation of the background, characterization is made for each region independently as a mixture of K Gaussian modes, considering the model of the approximation and detail coefficients at the different wavelet decomposition levels. Background region characterization is updated along time, and the detection of elements of interest is carried out computing the distance between background region models and those of each incoming image in the sequence. The inclusion of the context in the modeling scheme through each region characterization makes the model robust, being able to support not only gradual illumination and long-term changes, but also sudden illumination changes and the presence of strong shadows in the scene.

Object tracking through particle filtering has been widely addressed in recent years. However, most works assume a constant number of objects or utilize an external detector that monitors the entry or exit of objects in the scene. In this work, a novel tracking method based on particle filtering that is able to automatically track a variable number of objects is presented. As opposed to classical prior data assignment approaches, adaptation of tracks to the measurements is managed globally. Additionally, the designed particle filter is able to generate hypotheses on the presence of new objects in the scene, and to confirm or dismiss them by gradually adapting to the global observation. The method is especially suited for environments where traditional object detectors render noisy measurements and frequent artifacts, such as that given by a camera mounted on a vehicle, where it is proven to yield excellent results.

A new method for robust estimation of vanishing points is introduced in this paper. It is based on the MSAC (M-estimator Sample and Consensus) algorithm and on the definition of a new distance function between a vanishing point and a given orientation. Apart from the robustness, our method represents a flexible and efficient solution, since it allows to work with different type of image data, and its iterative nature makes better use of the available information to obtain more accurate estimates. The key issue of the work is the proposed distance function, that makes the error to be independent from the position of an hypothesized vanishing point, which allows to work with points at the infinity. Besides, the estimation process is guided by a non-linear optimization process that enhances the accuracy of the system. The robustness of our proposal, compared with other methods in the literature is shown with a set of tests carried out for both synthetic data and real images. The results show that our approach obtain excellent levels of accuracy and that is definitely robust against the presence of large amounts of outliers, outperforming other state of the art approaches.

This paper introduces a new strategy for plane rectification in sequences of images, based on the Expectation-Maximization (EM) algorithm. Our approach is able to compute simultaneously the parameters of the dominant vanishing point in the image plane and the most significant lines passing through it. It is based on a novel definition of the likelihood distribution of the gradient image considering both the position and the orientation of the gradient pixels. Besides, the mixture model in which the EM algorithm operates is extended, compared to other works, to consider an additional component to control the presence of outliers. Some synthetic data tests are described to show the robustness and efficiency of the proposed method. The plane rectification results show that the method is able to remove the perspective and affine distortion of real traffic sequences without the need to compute two vanishing points.

This paper presents a full system for vehicle detection and tracking in non-stationary settings based on computer vision. The method proposed for vehicle detection exploits the geometrical relations between the elements in the scene so that moving objects (i.e., vehicles) can be detected by analyzing motion parallax. Namely, the homography of the road plane between successive images is computed. Most remarkably, a novel probabilistic framework based on Kalman filtering is presented for reliable and accurate homography estimation. The estimated homography is used for image alignment, which in turn allows to detect the moving vehicles in the image. Tracking of vehicles is performed on the basis of a multidimensional particle filter, which also manages the exit and entries of objects. The filter involves a mixture likelihood model that allows a better adaptation of the particles to the observed measurements. The system is specially designed for highway environments, where it has been proven to yield excellent results.

Visual tracking with moving cameras is a challenging task. The global motion induced by the moving camera moves the target object outside the expected search area, according to the object dynamics. The typical approach is to use a registration algorithm to compensate the camera motion. However, in situations involving several moving objects, and backgrounds highly affected by the aperture problem, image registration quality may be very low, decreasing dramatically the performance of the tracking. In this work, a novel approach is proposed to successfully tackle the tracking with moving cameras in complex situations, which involve several independent moving objects. The key idea is to compute several hypotheses for the camera motion, instead of estimating deterministically only one. These hypotheses are combined with the object dynamics in a particle filter framework to predict the most probable object locations. Then, each hypothetical object location is evaluated by the measurement model using a spatiogram, which is a region descriptor based on color and spatial distributions. Experimental results show that the proposed strategy allows to accurately track an object in complex situations affected by strong ego motion.

Multiple object tracking is a main research area in the computer vision field. Particle filters have shown their performance as a powerful tool allowing to track visual objects giving temporal coherence to incoming observations, as well as offering an excellent framework for this task due to its inherent multimodality. However, traditional algorithms for particle filters do not cope directly with multiple objects and several considerations have to be addressed. In this work, an efficient reclustering strategy is proposed, which takes into account new measurements according to a novelty function, and provides a criterium to determine the minimum required number of particles to be drawn for each tracked object. To show its performance, this strategy has been used as a multiple 2D object tracking for video-surveillance applications. Excellent results are obtained, in terms of efficiency and accuracy.

This paper presents a robust method for real-time vehicle detection and tracking in dynamic traffic environments. The proposed strategy aims to find a trade-off between the robustness shown by time-uncorrelated detection techniques and the speed-up obtained with tracking algorithms. It combines both advantages by continuously evaluating the quality of the tracking results along time and triggering new detections to restart the tracking process when quality falls behind a certain quality requirement. Robustness is also ensured within the tracking algorithm with an outlier rejection stage and the use of stochastic filtering. Several sequences from real traffic situations have been tested, obtaining highly accurate multiple vehicle detections.

People positioning and tracking in 3D indoor environments are challenging tasks due to background clutter and occlusions. Current works are focused on solving people occlusions in low-cluttered backgrounds, but fail in high-cluttered scenarios, specially when foreground objects occlude people. In this paper, a novel 3D people positioning and tracking system is presented, which shows itself robust to both possible occlusion sources: static scene objects and other people. The system holds on a set of multiple cameras with partially overlapped fields of view. Moving regions are segmented independently in each camera stream by means of a new background modeling strategy based on Gabor filters. People detection is carried out on these segmentations through a template-based correlation strategy. Detected people are tracked independently in each camera view by means of a graph-based matching strategy, which estimates the best correspondences between consecutive people segmentations. Finally, 3D tracking and positioning of people is achieved by geometrical consistency analysis over the tracked 2D candidates, using head position (instead of object centroids) to increase robustness to foreground occlusions.

Road modeling is the first step towards environment perception within driver assistance video-based systems. Typically, lane modeling allows applications such as lane departure warning or lane invasion by other vehicles. In this paper, a new monocular image processing strategy that achieves a robust multiple lane model is proposed. The identification of multiple lanes is done by firstly detecting the own lane and estimating its geometry under perspective distortion. The perspective analysis and curve fitting allows to hypothesize adjacent lanes assuming some a priori knowledge about the road. The verification of these hypotheses is carried out by a confidence level analysis. Several types of sequences have been tested, with different illumination conditions, presence of shadows and significant curvature, all performing in realtime. Results show the robustness of the system, delivering accurate multiple lane road models in most situations.

This paper presents a new in-vehicle real-time vehicle detection strategy which hypothesizes the presence of vehicles in rectangular sub-regions based on the robust classification of features vectors result of a combination of multiple morphological vehicle features. One vector is extracted for each region of the image likely containing vehicles as a multidimensional likelihood measure with respect to a simplified vehicle model. A supervised training phase set the representative vectors of the classes vehicle and non-vehicle, so that the hypothesis is verified or not according to the Mahalanobis distance between the feature vector and the representative vectors. Excellent results have been obtained in several video sequences accurately detecting vehicles with very different aspect-ratio, color, size, etc, while minimizing the number of missing detections and false alarms.

An efficient automatic moving target detection and tracking system in airborne forward looking infrared (FLIR) imagery is presented in this paper. Due to camera ego-motion, these detection and tracking tasks are challenging problems. Besides, previously proposed techniques are not suitable for aerial images, as the predominant regions are non-textured. The proposed system efficiently estimates not only the camera motion but also the target motion, by means of an accurate motion vector field computation and robust motion parameters estimation technique. This information allows accurately to segment each target, and tracking them with ego-motion compensation. Verification of tracking restrictions helps detecting true targets while reducing very significantly the false alarm rate. Excellent results have been obtained over real FLIR sequences.

In this paper we propose a new fast mode decision (FMD) algorithm to reduce the computational load of the motion estimation (ME) process of the new video coding standard H.264/AVC main profile. The algorithm firstly computes the skip/direct mode (respectively for P and B-frames), for all macroblocks (MBs), and then decides, for each MB if no mores modes are needed. This decision is based on the generation of predictions of the expected PSNR of the frame to be encoded. These predictions are performed with the distortion cost obtained for skip/direct mode jointly with the average distortion cost obtained for the previous encoded frames. Results show that the computational load have been dramatically reduced, with average encoding time saving of-36.38%, while nearly negligible loss of coding efficiency.

In this paper we propose a new fast mode decision (FMD) algorithm for H.264/AVC to reduce the computational load of the motion estimation (ME) process. It is oriented to dramatically reduce the encoding time regardless of the level of motion present at the sequences (high or low-motion). The algorithm decides as the best mode the skip mode or mode 1 without need to compute the rest of coding alternatives if some conditions are satisfied. These conditions ensure these modes to be used appropriately for low motion and high motion sequences respectively to achieve an appropriate rate distortion (RD) cost based on the results of the previous encoded frames. Tests have shown reductions of the encoding time around-81% for all kinds of sequences, while moderate loss of coding efficiency.

A new adaptive segmentation strategy is proposed to segment gymnasts in sport sequences accurately. It is based on a Markov random fields (MRF) change detection analysis operating on a multiresolution combination of static and dynamic image differences. After a morphological analysis of the segmented masks, estimated motion information in the area of interest is incorporated to improve the efficiency of the segmentation process. Although presented in the particular context of gymnastic exercises, the new segmentation strategy could be applied to other applications where moving objects on a quasi-static background need to be segmented.

An efficient segmentation technique of regular shaped sudden temporal intensity changes in image sequences is presented. It is based on morphological analysis of binarized difference images, and its parameters are dynamically modified according to an efficient estimation of the global motion of the scene. Motion analysis is carried out with the combination of a multiresolution flow estimation technique and a closed formula over the resulting motion vectors based on a simplified motion model. Specially suited for applications demanding a temporal reference, the new segmentation technique will be described in the context of a time stamp detection application.

A new segmentation and analysis strategy to automatically measure the inhibition halos (IHs) in antibiograms is presented. It is based on the application of a combined texture and directional filtering analysis technique on the result of a segmentation process which incorporates colour analysis and Hough transform applications. The computational efficiency is highly improved by applying the different processing stages to selected areas of the acquired images and exploiting a priori information about the IHs shape and distribution.

Motion description is a key clement in region-based video-coding approaches, becoming essential for low and very low bit-rate transmissions. Among the different advanced motion description approaches, multivector motion description (MMD) models the region motion through a variable number of motion vectors (MVs) applied on specific region points called control points (CPs). However, its application for region-based video coding would require the transmission of both MVs and CPs in order to adequately compensate each region motion. Here, a region-based control point determination strategy for MMD is proposed, which is strictly based on region contour information and the application of motion-interpolation accuracy constraints to drive the CP's location procedure. This approach offers two main advantages while keeping a reasonable computational cost: (1) it is able to operate on arbitrary-shaped regions of the image and (2) it allows the complete elimination of the transmission of any information related to the number and position of the CPs.

The improvement of vehicle security is a major priority of the car industry. Both active and passive security-systems have experienced a great development during the last decade, but the main research is still focused on minimising the errors committed by the driver rather than trying to avoid them. In this paper, a basic monocular vision system, focused on road location solely, is described. Working with only one video camera hinders the exact 3D reconstruction of the scene: no information about distances and dimensions is available, unless a priori artificial constraints are taken. The problem increases because of the movement of the camera with respect to the scene. In the present system, the road searching and following algorithms operate on the two-dimensional image plane, and 2D to 3D conversion is not accomplished. The system obtains excellent final results, succeeding in more than 75% of the analysed images from the tested sequences.

An efficient strategy to determine the prediction error regions to be coded within a region-based prediction error coding scheme is presented. Prediction error coding is based on the segmentation of the displaced field difference (DFD) and coding the resulting arbitrary shaped DFD regions using shape adaptive DCT. Efficiency in the determination of the DFD regions to be coded is achieved by eliminating from the selection process the direct computation of the cost of region contours and textures coding. With this scheme, perceptual distortion of the decoded images is reduced while quality is locally improved on relevant image areas. Comparative results with the complete H.263 coder are shown.

An intra-frame segmentation strategy to assist region-based motion estimation and compensation is presented. It is based on the multiresolution application of a histogram clustering and a probabilistic relaxation-labeling algorithm, followed by a local gradient-based bottom-up merging procedure. Specially suited for region-based video coding, it strongly differs from other proposals in that it generates arbitrary shaped image regions with pixel accuracy at a low computational cost, while allowing full reconstruction of the segmentation at the decoder without the transmission of any region description information.

An intra-frame segmentation strategy to assist region-based motion estimation and compensation is presented. It is based on the multiresolution application of a histogram clustering and a probabilistic relaxation labelling algorithm, followed by a focal gradient-based bottom-up merging procedure. Specially suited for region-based video coding, it strongly differs from other proposals in that it generates arbitrary shaped image regions with pixel accuracy at a low computational cost, while allowing full reconstruction of the segmentation at the decoder without the transmission of any region description information.

In spite of their advantages, region-based approaches present a heavy burden in the required amount of information to describe region contours. To avoid these limitations, a new region-based motion estimation and compensation strategy is proposed, which allows the operation on arbitrary shaped regions and the ability to reconstruct them without any contour information. The strategy is enhanced by the use of multivector motion estimation and compensation. Results showing the quality and advantages of the strategy are provided.

A new method for the determination of adaptive palettes of color images with a variation of Kohonen (1984) networks, for the purposes of image storage and display, is presented. New neighborhood relationships in one and three dimensions are defined among neurons and the scope of these binds is also extended. The presented method is compared with popular palette selection algorithms, the resulting quality, provided network convergence, is competitive with the other methods, both for natural color and graphic images. Additionally, the resulting palettes show interesting topological relationships among colors.

An analytical procedure for the estimation of the focus of expansion (FOE) location is introduced. The method is applied to a vector field that has been obtained from the analysis of the relative translational movement of a rigid body with respect to the acquiring camera (monocular system). Two basic properties of complex differentiation theory are applied, assuming that each vector of the optic flow is a real-valued complex function of the form v(x,y)=v/sub x/(x,y)+iv/sub y/,(x,y). These properties provide the set of points for which both components of the vector field v/sub x/(x,y) and v/sub y/(x,y) are harmonic and, as a subset, the set of points for which both components are null. The use of perspective projection states that the FOE is obtained as the locus where v/sup x/(x,y)=0=v/sub y/(x,y) and, therefore, where the conditions stated by both properties meet.

Pyramidal encoding schemes are proposed as an approximation to a subband coding approach into a packet switched transmission environment. In this paper, the method is introduced and justified, extending the classic approach to pyramidal encoding with a pruning mechanism which reduces extensively (in the nearabouts of 3:1) the information to be coded with a very limited quality loss. The different parameters associated to the coding structure are studied, very particularly the behaviour of the input buffer where the subbands are implemented and important delays have to be accounted for. Results are supplied showing the most meaningful design implications that have been found. A detailed description of the simulator which has been developed is also given, expressing with detail the way of handling the asynchrony of the overall process.