Supervised training of neural networks requires a large amount of manually annotated data and the resulting networks tend to be sensitive to out-of-distribution (OOD) data. Self- and semi-supervised training schemes reduce the amount of annotated data required during the training process. However, OOD generalization remains a major challenge for most methods. Strategies that promote smoother decision boundaries play an important role in out-of-distribution generalization. For example, embedding propagation (EP) for manifold smoothing has recently shown to considerably improve the OOD performance for few-shot classification. EP achieves smoother class manifolds by building a graph from sample embeddings and propagating information through the nodes in an unsupervised manner. In this work, we extend the original EP paper providing additional evidence and experiments showing that it attains smoother class embedding manifolds and improves results in settings beyond few-shot classification. Concretely, we show that EP improves the robustness of neural networks against multiple adversarial attacks as well as semi- and self-supervised learning performance.

Pain is an unpleasant feeling that has been shown to be an important factor for the recovery of patients. Since this is costly in human resources and difficult to do objectively, there is the need for automatic systems to measure it. In this paper, contrary to current state-of-the-art techniques in pain assessment, which are based on facial features only, we suggest that the performance can be enhanced by feeding the raw frames to deep learning models, outperforming the latest state-of-the-art results while also directly facing the problem of imbalanced data. As a baseline, our approach first uses convolutional neural networks (CNNs) to learn facial features from VGG_Faces, which are then linked to a long short-term memory to exploit the temporal relation between video frames. We further compare the performances of using the so popular schema based on the canonically normalized appearance versus taking into account the whole image. As a result, we outperform current state-of-the-art area under the curve performance in the UNBC-McMaster Shoulder Pain Expression Archive Database. In addition, to evaluate the generalization properties of our proposed methodology on facial motion recognition, we also report competitive results in the Cohn Kanade+ facial expression database.

Fine-grained image recognition is central to many multimedia tasks such as search, retrieval, and captioning. Unfortunately, these tasks are still challenging since the appearance of samples of the same class can be more different than those from different classes. This issue is mainly due to changes in deformation, pose, and the presence of clutter. In the literature, attention has been one of the most successful strategies to handle the aforementioned problems. Attention has been typically implemented in neural networks by selecting the most informative regions of the image that improve classification. In contrast, in this paper, attention is not applied at the image level but to the convolutional feature activations. In essence, with our approach, the neural model learns to attend to lower-level feature activations without requiring part annotations and uses those activations to update and rectify the output likelihood distribution. The proposed mechanism is modular, architecture-independent, and efficient in terms of both parameters and computation required. Experiments demonstrate that well-known networks such as wide residual networks and ResNeXt, when augmented with our approach, systematically improve their classification accuracy and become more robust to changes in deformation and pose and to the presence of clutter. As a result, our proposal reaches state-of-the-art classification accuracies in CIFAR-10, the Adience gender recognition task, Stanford Dogs, and UEC-Food100 while obtaining competitive performance in ImageNet, CIFAR-100, CUB200 Birds, and Stanford Cars. In addition, we analyze the different components of our model, showing that the proposed attention modules succeed in finding the most discriminative regions of the image. Finally, as a proof of concept, we demonstrate that with only local predictions, an augmented neural network can successfully classify an image before reaching any fully connected layer, thus reducing the computational amount up to 10%.

Most advanced driving assistance systems already include pedestrian detection systems. Unfortunately, there is still a tradeoff between precision and real time. For a reliable detection, excellent precision-recall such a tradeoff is needed to detect as many pedestrians as possible while, at the same time, avoiding too many false alarms; in addition, a very fast computation is needed for fast reactions to dangerous situations. Recently, novel approaches based on deformable templates have been proposed since these show a reasonable detection performance although they are computationally too expensive for real-time performance. In this paper, we present a system for pedestrian detection based on a hierarchical multiresolution part-based model. The proposed system is able to achieve state-of-the-art detection accuracy due to the local deformations of the parts while exhibiting a speedup of more than one order of magnitude due to a fast coarse-to-fine inference technique. Moreover, our system explicitly infers the level of resolution available so that the detection of small examples is feasible with a very reduced computational cost. We conclude this contribution by presenting how a graphics processing unit-optimized implementation of our proposed system is suitable for real-time pedestrian detection in terms of both accuracy and speed.

Abstract This paper describes an approach to human action recognition based on a probabilistic optimization model of body parts using hidden Markov model (HMM). Our method is able to distinguish between similar actions by only considering the body parts having major contribution to the actions, for example, legs for walking, jogging and running; arms for boxing, waving and clapping. We apply HMMs to model the stochastic movement of the body parts for action recognition. The HMM construction uses an ensemble of body‐part detectors, followed by grouping of part detections, to perform human identification. Three example‐based body‐part detectors are trained to detect three components of the human body: the head, legs and arms. These detectors cope with viewpoint changes and self‐occlusions through the use of ten sub‐classifiers that detect body parts over a specific range of viewpoints. Each sub‐classifier is a support vector machine trained on features selected for the discriminative power for each particular part/viewpoint combination. Grouping of these detections is performed using a simple geometric constraint model that yields a viewpoint‐invariant human detector. We test our approach on three publicly available action datasets: the KTH dataset, Weizmann dataset and HumanEva dataset. Our results illustrate that with a simple and compact representation we can achieve robust recognition of human actions comparable to the most complex, state‐of‐the‐art methods.

Recent progress in the field of human action recognition points towards the use of Spatio-Temporal Interest Points (STIPs) for local descriptor-based recognition strategies. In this paper we present a new approach for STIP detection by applying surround suppression combined with local and temporal constraints. Our method is significantly different from existing STIP detectors and improves the performance by detecting more repeatable, stable and distinctive STIPs for human actors, while suppressing unwanted background STIPs. For action representation we use a bag-of-visual words (BoV) model of local N-jet features to build a vocabulary of visual-words. To this end, we introduce a novel vocabulary building strategy by combining spatial pyramid and vocabulary compression techniques, resulting in improved performance and efficiency. Action class specific Support Vector Machine (SVM) classifiers are trained for categorization of human actions. A comprehensive set of experiments on existing benchmark datasets, and more challenging datasets of complex scenes, validate our approach and show state-of-the-art performance.

In this paper we describe a novel approach to reactive tracking of moving targets with a pan-tilt-zoom camera. The approach uses an extended Kalman filter to jointly track the object position in the real world, its velocity in 3D and the camera intrinsics, in addition to the rate of change of these parameters. The filter outputs are used as inputs to PID controllers which continuously adjust the camera motion in order to reactively track the object at a constant image velocity while simultaneously maintaining a desirable target scale in the image plane. We provide experimental results on simulated and real tracking sequences to show how our tracker is able to accurately estimate both 3D object position and camera intrinsics with very high precision over a wide range of focal lengths.

In this paper we propose a method that smartly improves occlusion handling in stereo matching using trinocular stereo. The main idea is based on the assumption that any occluded region in a matched stereo pair (middle-left images) in general is not occluded in the opposite matched pair (middle-right images). Then two disparity space images (DSI) are merged in one composite DSI. The proposed integration differs from the known approach that uses a cumulative cost. The experimental results are evaluated on the Middlebury data set, showing high performance of the proposed algorithm especially in the occluded regions. Our method solves the problem on the base of a real matching cost, in such a way a global optimization problem is solved just once, and the resultant solution does not have to be corrected in the occluded regions. In contrast, the traditional methods that use two images approach have to complicate a lot their algorithms by additional add hog or heuristic techniques to reach competitive results in occluded regions.

Abstract In this paper we present a novel approach to generate augmented video sequences in real‐time, involving interactions between virtual and real agents in real scenarios. On the one hand, real agent motion is estimated by means of a multi‐object tracking algorithm, which determines real objects' position over the scenario for each time step. On the other hand, virtual agents are provided with behavior models considering their interaction with the environment and with other agents. The resulting framework allows to generate video sequences involving behavior‐based virtual agents that react to real agent behavior and has applications in education, simulation, and in the game and movie industries. We show the performance of the proposed approach in an indoor and outdoor scenario simulating human and vehicle agents. Copyright © 2009 John Wiley & Sons, Ltd.

In this paper, we address the recognition of subtle facial expressions by reasoning on the classification confidence. Psychological evidences have determined that eyelids and eyebrows are significant for the recognition of subtle facial expressions and the early perception of human emotions. This early perception results in a more complex problem, which requires a confidence assessment for any provided solution. Thus, traditional score-based classifiers (e.g. k-NN and NN) are not able to produce confident estimates. Instead, we first present five confidence estimators and a confidence classification assessment for Case-Based Reasoning (CBR). Second, we improve the expression retrieval from the database by learning the neighbourhood's dimensions for the expected classification confidences. Third, we reuse the previous classified expressions and the confidence assessment to improve the classification achieved by k-NN. Fourth, we improve the database for generalization with new subjects by learning thresholds to minimize misclassification with low confidence, maximize correct classifications with high confidence and re-arrange misclassification with high confidence. The proposed system represents an effective contribution for both subtle expression recognition and CBR methodology. It achieves an average recognition of 97% plusmn 1% with a confidence of 96% plusmn 2% for expressiveness between 20% and 100%.

Real applications on people tracking are usually based on image heuristics. Real approaches do not use to apply recent prediction-estimation theoretical frameworks. These require the definition of complex dynamical and shape object models before the tracking process. We present a probablistic framework that takes profit of these theories adapting them to real applications. The key idea of this work is to estimate the shape model and dynamical objects parameters using only image data. The flexibility of our algorithm makes it suitable to be used on different real applications. Some experiments have been done in order to test our method in outdoor scenes people tracking.