Abstract: Semantic segmentation refers to the process of assigning an object label (e.g., building, road, sidewalk, car, pedestrian) to every pixel in an image. Common approaches formulate the task as a random field labeling problem modeling the interactions between labels by combining local and contextual features such as color, depth, edges, SIFT or HoG. These models are trained to maximize the likelihood of the correct classification given a training set. However, these approaches rely on hand–designed features (e.g., texture, SIFT or HoG) and a higher computational time required in the inference process.
Therefore, in this paper, we focus on estimating the unary potentials of a conditional random field via ensembles of learned features. We propose an algorithm based on convolutional neural networks to learn local features from training data at different scales and resolutions. Then, diversification between these features is exploited using a weighted linear combination. Experiments on a publicly available database show the effectiveness of the proposed method to perform semantic road scene segmentation in still images. The algorithm outperforms appearance based methods and its performance is similar compared to state–of–the–art methods using other sources of information such as depth, motion or stereo.

Abstract: Project-Based Learning (PBL) is an educational strategy to improve student’s learning capability that, in recent years, has had a progressive acceptance in undergraduate studies. This methodology is based on solving a problem or project in a student working group. In this way, PBL focuses on learning the necessary tools to correctly find a solution to given problems. Since the learning initiative is transferred to the student, the PBL method promotes students own abilities. This allows a better assessment of the true workload that carries out the student in the subject. It follows that the methodology conforms to the guidelines of the Bologna document, which quantifies the student workload in a subject by means of the European credit transfer system (ECTS). PBL is currently applied in undergraduate studies needing strong practical training such as medicine, nursing or law sciences. Although this is also the case in engineering studies, amazingly, few experiences have been reported. In this paper we propose to use PBL in the educational organization of the Computer Graphics subjects in the Computer Science degree. Our PBL project focuses in the development of a C++ graphical environment based on the OpenGL libraries for visualization and handling of different graphical objects. The starting point is a basic skeleton that already includes lighting functions, perspective projection with mouse interaction to change the point of view and three predefined objects. Students have to complete this skeleton by adding their own functions to solve the project. A total number of 10 projects have been proposed and successfully solved. The exercises range from human face rendering to articulated objects, such as robot arms or puppets. In the present paper we extensively report the statement and educational objectives for two of the projects: solar system visualization and a chess game. We report our earlier educational experience based on the standard classroom theoretical, problem and practice sessions and the reasons that motivated searching for other learning methods. We have mainly chosen PBL because it improves the student learning initiative. We have applied the PBL educational model since the beginning of the second semester. The student’s feedback increases in his interest for the subject. We present a comparative study of the teachers’ and students’ workload between PBL and the classic teaching approach, which suggests that the workload increase in PBL is not as high as it seems.

Abstract: State-of-the-art deformable part-based models based on latent SVM have shown excellent results on human detection. In this paper, we propose to train a multiview deformable part-based model with automatically generated part examples from virtual-world data. The method is efficient as: (i) the part detectors are trained with precisely extracted virtual examples, thus no latent learning is needed, (ii) the multiview pedestrian detector enhances the performance of the pedestrian root model, (iii) a top-down approach is used for part detection which reduces the searching space. We evaluate our model on Daimler and Karlsruhe Pedestrian Benchmarks with publicly available Caltech pedestrian detection evaluation framework and the result outperforms the state-of-the-art latent SVM V4.0, on both average miss rate and speed (our detector is ten times faster).

Abstract: Image based human detection is of paramount interest due to its potential applications in fields such as advanced driving assistance, surveillance and media analysis. However, even detecting non-occluded standing humans remains a challenge of intensive research. The most promising human detectors rely on classifiers developed in the discriminative paradigm, i.e., trained with labelled samples. However, labeling is a manual intensive step, especially in cases like human detection where it is necessary to provide at least bounding boxes framing the humans for training. To overcome such problem, some authors have proposed the use of a virtual world where the labels of the different objects are obtained automatically. This means that the human models (classifiers) are learnt using the appearance of rendered images, i.e., using realistic computer graphics. Later, these models are used for human detection in images of the real world. The results of this technique are surprisingly good. However, these are not always as good as the classical approach of training and testing with data coming from the same camera, or similar ones. Accordingly, in this paper we address the challenge of using a virtual world for gathering (while playing a videogame) a large amount of automatically labelled samples (virtual humans and background) and then training a classifier that performs equal, in real-world images, than the one obtained by equally training from manually labelled real-world samples. For doing that, we cast the problem as one of domain adaptation. In doing so, we assume that a small amount of manually labelled samples from real-world images is required. To collect these labelled samples we propose a non-standard active learning technique. Therefore, ultimately our human model is learnt by the combination of virtual and real world labelled samples (Fig. 1), which has not been done before. We present quantitative results showing that this approach is valid.

Abstract: Vision-based object detectors are crucial for different applications. They rely on learnt object models. Ideally, we would like to deploy our vision system in the scenario where it must operate, and lead it to self-learn how to distinguish the objects of interest, i.e., without human intervention. However, the learning of each object model requires labelled samples collected through a tiresome manual process. For instance, we are interested in exploring the self-training of a pedestrian detector for driver assistance systems. Our first approach to avoid manual labelling consisted in the use of samples coming from realistic computer graphics, so that their labels are automatically available [12]. This would make possible the desired self-training of our pedestrian detector. However, as we showed in [14], between virtual and real worlds it may be a dataset shift. In order to overcome it, we propose the use of unsupervised domain adaptation techniques that avoid human intervention during the adaptation process. In particular, this paper explores the use of the transductive SVM (T-SVM) learning algorithm in order to adapt virtual and real worlds for pedestrian detection (Fig. 1).