Daniel Ponsa

Virtual and Real World Adaptationfor Pedestrian Detection

Pedestrian detection is of paramount interest for many applications. Most promising detectors rely on discriminatively learnt classifiers, i.e., trained with annotated samples. However, the annotation step is a human intensive and subjective task worth to be minimized. By using virtual worlds we can automatically obtain precise and rich annotations. Thus, we face the question: can a pedestrian appearance model learnt in realistic virtual worlds work successfully for pedestrian detection in real-world images? Conducted experiments show that virtual-world based training can provide excellent testing accuracy in real world, but it can also suffer the data set shift problem as real-world based training does. Accordingly, we have designed a domain adaptation framework, V-AYLA, in which we have tested different techniques to collect a few pedestrian samples from the target domain (real world) and combine them with the many examples of the source domain (virtual world) in order to train a domain adapted pedestrian classifier that will operate in the target domain. V-AYLA reports the same detection accuracy than when training with many human-provided pedestrian annotations and testing with real-world images of the same domain. To the best of our knowledge, this is the first work demonstrating adaptation of virtual and real worlds for developing an object detector.

2D-3D-based on-board pedestrian detection system

During the next decade, on-board pedestrian detection systems will play a key role in the challenge of increasing traffic safety. The main target of these systems, to detect pedestrians in urban scenarios, implies overcoming difficulties like processing outdoor scenes from a mobile platform and searching for aspect-changing objects in cluttered environments. This makes such systems combine techniques in the state-of-the-art Computer Vision. In this paper we present a three module system based on both 2D and 3D cues. The first module uses 3D information to estimate the road plane parameters and thus select a coherent set of regions of interest (ROIs) to be further analyzed. The second module uses Real AdaBoost and a combined set of Haar wavelets and edge orientation histograms to classify the incoming ROIs as pedestrian or non-pedestrian. The final module loops again with the 3D cue in order to verify the classified ROIs and with the 2D in order to refine the final results. According to the results, the integration of the proposed techniques gives rise to a promising system.

3D vehicle sensor based on monocular vision

Determining the position of other vehicles on the road is a key information to help driver assistance systems to increase drivers safety. Accordingly, the work presented in this paper addresses the problem of detecting the vehicles in front of our own one and estimating their 3D position by using a single monochrome camera. Rather than using predefined high level image features as symmetry, shadow search, etc., our proposal for the vehicle detection is based on a learning process that determines, from a training set, which are the best features to distinguish vehicles from non-vehicles. To compute 3D information with a single camera a key point consists of knowing the position where the horizon projects onto the image. However, this position can change in every frame and is difficult to determine. In this paper we study the coupling between the perceived horizon and the actual width of vehicles in order to reduce the uncertainty in their estimated 3D position derived from an unknown horizon.

Haar Wavelets and Edge Orientation Histograms for On---Board Pedestrian Detection

On---board pedestrian detection is a key task in advanced driver assistance systems. It involves dealing with aspect---changing objects in cluttered environments, and working in a wide range of distances, and often relies on a classification step that labels image regions of interest as pedestrians or non---pedestrians. The performance of this classifier is a crucial issue since it represents the most important part of the detection system, thus building a good classifier in terms of false alarms, missdetection rate and processing time is decisive. In this paper, a pedestrian classifier based on Haar wavelets and edge orientation histograms (HW+EOH) with AdaBoost is compared with the current state---of---the---art best human---based classifier: support vector machines using histograms of oriented gradients (HOG). The results show that HW+EOH classifier achieves comparable false alarms/missdetections tradeoffs but at much lower processing time than HOG.

An Efficient Approach to Onboard Stereo Vision System Pose Estimation

This paper presents an efficient technique for estimating the pose of an onboard stereo vision system relative to the environments dominant surface area, which is supposed to be the road surface. Unlike previous approaches, it can be used either for urban or highway scenarios since it is not based on a specific visual traffic feature extraction but on 3D raw data points. The whole process is performed in the Euclidean space and consists of two stages. Initially, a compact 2D representation of the original 3D data points is computed. Then, a RANdom SAmple Consensus (RANSAC) based least-squares approach is used to fit a plane to the road. Fast RANSAC fitting is obtained by selecting points according to a probability function that takes into account the density of points at a given depth. Finally, stereo camera height and pitch angle are computed related to the fitted road plane. The proposed technique is intended to be used in driver-assistance systems for applications such as vehicle or pedestrian detection. Experimental results on urban environments, which are the most challenging scenarios (i.e., flat/uphill/downhill driving, speed bumps, and cars accelerations), are presented. These results are validated with manually annotated ground truth. Additionally, comparisons with previous works are presented to show the improvements in the central processing unit processing time, as well as in the accuracy of the obtained results.

Learning a Part-Based Pedestrian Detector in a Virtual World

Detecting pedestrians with on-board vision systems is of paramount interest for assisting drivers to prevent vehicle-to-pedestrian accidents. The core of a pedestrian detector is its classification module, which aims at deciding if a given image window contains a pedestrian. Given the difficulty of this task, many classifiers have been proposed during the last 15 years. Among them, the so-called (deformable) part-based classifiers, including multiview modeling, are usually top ranked in accuracy. Training such classifiers is not trivial since a proper aspect clustering and spatial part alignment of the pedestrian training samples are crucial for obtaining an accurate classifier. In this paper, we first perform automatic aspect clustering and part alignment by using virtual-world pedestrians, i.e., human annotations are not required. Second, we use a mixture-of-parts approach that allows part sharing among different aspects. Third, these proposals are integrated in a learning framework, which also allows incorporating real-world training data to perform domain adaptation between virtual- and real-world cameras. Overall, the obtained results on four popular on-board data sets show that our proposal clearly outperforms the state-of-the-art deformable part-based detector known as latent support vector machine.

Video Alignment for Change Detection

In this work, we address the problem of aligning two video sequences. Such alignment refers to synchronization, i.e., the establishment of temporal correspondence between frames of the first and second video, followed by spatial registration of all the temporally corresponding frames. Video synchronization and alignment have been attempted before, but most often in the relatively simple cases of fixed or rigidly attached cameras and simultaneous acquisition. In addition, restrictive assumptions have been applied, including linear time correspondence or the knowledge of the complete trajectories of corresponding scene points; to some extent, these assumptions limit the practical applicability of any solutions developed. We intend to solve the more general problem of aligning video sequences recorded by independently moving cameras that follow similar trajectories, based only on the fusion of image intensity and GPS information. The novelty of our approach is to pose the synchronization as a MAP inference problem on a Bayesian network including the observations from these two sensor types, which have been proved complementary. Alignment results are presented in the context of videos recorded from vehicles driving along the same track at different times, for different road types. In addition, we explore two applications of the proposed video alignment method, both based on change detection between aligned videos. One is the detection of vehicles, which could be of use in ADAS. The other is online difference spotting videos of surveillance rounds.

Learning local feature descriptors with triplets and shallow convolutional neural networks.

It has recently been demonstrated that local feature descriptors based on convolutional neural networks (CNN) can significantly improve the matching performance. Previous work on learning such descriptors has focused on exploiting pairs of positive and negative patches to learn discriminative CNN representations. In this work, we propose to utilize triplets of training samples, together with in-triplet mining of hard negatives. We show that our method achieves state of the art results, without the computational overhead typically associated with mining of negatives and with lower complexity of the network architecture. We compare our approach to recently introduced convolutional local feature descriptors, and demonstrate the advantages of the proposed methods in terms of performance and speed. We also examine different loss functions associated with triplets.

Cascade of classifiers for vehicle detection

Being aware of other vehicles on the road ahead is a key information to help driver assistance systems to increase drivers safety. This paper addresses this problem, proposing a system to detect vehicles from the images provided by a single camera mounted in a mobile platform. A classifier-based approach is presented, based on the evaluation of a cascade of classifiers (COC) at different scanned image regions. The Adaboost algorithm is used to determine the COC from training sets. Two proposals are done to reduce the computation needed for the detection scheme used: a lazy evaluation of the COC, and the customization of the COC by a wrapping process. The benefits of these two proposals are quantified in terms of the average number of image features required to classify an image region, achieving a reduction of the 58% on this concept, while scarcely penalizing the detection accuracy of the system.

On-board image-based vehicle detection and tracking

In this paper we present a computer vision system for daytime vehicle detection and localization, an essential step in the development of several types of advanced driver assistance systems. It has a reduced processing time and high accuracy thanks to the combination of vehicle detection with lane-markings estimation and temporal tracking of both vehicles and lane markings. Concerning vehicle detection, our main contribution is a frame scanning process that inspects images according to the geometry of image formation, and with an Adaboost-based detector that is robust to the variability in the different vehicle types (car, van, truck) and lighting conditions. In addition, we propose a new method to estimate the most likely three-dimensional locations of vehicles on the road ahead. With regards to the lane-markings estimation component, we have two main contributions. First, we employ a different image feature to the other commonly used edges: we use ridges, which are better suited to this problem. Second, we adapt RANSAC, a generic robust estimation method, to fit a parametric model of a pair of lane markings to the image features. We qualitatively assess our vehicle detection system in sequences captured on several road types and under very different lighting conditions. The processed videos are available on a web page associated with this paper. A quantitative evaluation of the system has shown quite accurate results (a low number of false positives and negatives) at a reasonable computation time.