Rui Caseiro

High-Speed Tracking with Kernelized Correlation Filters

The core component of most modern trackers is a discriminative classifier, tasked with distinguishing between the target and the surrounding environment. To cope with natural image changes, this classifier is typically trained with translated and scaled sample patches. Such sets of samples are riddled with redundancies—any overlapping pixels are constrained to be the same. Based on this simple observation, we propose an analytic model for datasets of thousands of translated patches. By showing that the resulting data matrix is circulant, we can diagonalize it with the discrete Fourier transform, reducing both storage and computation by several orders of magnitude. Interestingly, for linear regression our formulation is equivalent to a correlation filter, used by some of the fastest competitive trackers. For kernel regression, however, we derive a new kernelized correlation filter (KCF), that unlike other kernel algorithms has the exact same complexity as its linear counterpart. Building on it, we also propose a fast multi-channel extension of linear correlation filters, via a linear kernel, which we call dual correlation filter (DCF). Both KCF and DCF outperform top-ranking trackers such as Struck or TLD on a 50 videos benchmark, despite running at hundreds of frames-per-second, and being implemented in a few lines of code (Algorithm 1). To encourage further developments, our tracking framework was made open-source.

Exploiting the circulant structure of tracking-by-detection with kernels

Recent years have seen greater interest in the use of discriminative classifiers in tracking systems, owing to their success in object detection. They are trained online with samples collected during tracking. Unfortunately, the potentially large number of samples becomes a computational burden, which directly conflicts with real-time requirements. On the other hand, limiting the samples may sacrifice performance. Interestingly, we observed that, as we add more and more samples, the problem acquires circulant structure. Using the well-established theory of Circulant matrices, we provide a link to Fourier analysis that opens up the possibility of extremely fast learning and detection with the Fast Fourier Transform. This can be done in the dual space of kernel machines as fast as with linear classifiers. We derive closed-form solutions for training and detection with several types of kernels, including the popular Gaussian and polynomial kernels. The resulting tracker achieves performance competitive with the state-of-the-art, can be implemented with only a few lines of code and runs at hundreds of frames-per-second. MATLAB code is provided in the paper (see Algorithm 1).

Semantic segmentation with second-order pooling

Feature extraction, coding and pooling, are important components on many contemporary object recognition paradigms. In this paper we explore novel pooling techniques that encode the second-order statistics of local descriptors inside a region. To achieve this effect, we introduce multiplicative second-order analogues of average and max-pooling that together with appropriate non-linearities lead to state-of-the-art performance on free-form region recognition, without any type of feature coding. Instead of coding, we found that enriching local descriptors with additional image information leads to large performance gains, especially in conjunction with the proposed pooling methodology. We show that second-order pooling over free-form regions produces results superior to those of the winning systems in the Pascal VOC 2011 semantic segmentation challenge, with models that are 20,000 times faster.

Globally optimal solution to multi-object tracking with merged measurements

Multiple object tracking has been formulated recently as a global optimization problem, and solved efficiently with optimal methods such as the Hungarian Algorithm. A severe limitation is the inability to model multiple objects that are merged into a single measurement, and track them as a group, while retaining optimality. This work presents a new graph structure that encodes these multiple-match events as standard one-to-one matches, allowing computation of the solution in polynomial time. Since identities are lost when objects merge, an efficient method to identify groups is also presented, as a flow circulation problem. The problem of tracking individual objects across groups is then posed as a standard optimal assignment. Experiments show increased performance on the PETS 2006 and 2009 datasets compared to state-of-the-art algorithms.

Beyond Hard Negative Mining: Efficient Detector Learning via Block-Circulant Decomposition

Competitive sliding window detectors require vast training sets. Since a pool of natural images provides a nearly endless supply of negative samples, in the form of patches at different scales and locations, training with all the available data is considered impractical. A staple of current approaches is hard negative mining, a method of selecting relevant samples, which is nevertheless expensive. Given that samples at slightly different locations have overlapping support, there seems to be an enormous amount of duplicated work. It is natural, then, to ask whether these redundancies can be eliminated. In this paper, we show that the Gram matrix describing such data is block-circulant. We derive a transformation based on the Fourier transform that block-diagonalizes the Gram matrix, at once eliminating redundancies and partitioning the learning problem. This decomposition is valid for any dense features and several learning algorithms, and takes full advantage of modern parallel architectures. Surprisingly, it allows training with all the potential samples in sets of thousands of images. By considering the full set, we generate in a single shot the optimal solution, which is usually obtained only after several rounds of hard negative mining. We report speed gains on Caltech Pedestrians and INRIA Pedestrians of over an order of magnitude, allowing training on a desktop computer in a couple of minutes.

Free-Form Region Description with Second-Order Pooling

Semantic segmentation and object detection are nowadays dominated by methods operating on regions obtained as a result of a bottom-up grouping process (segmentation) but use feature extractors developed for recognition on fixed-form (e.g. rectangular) patches, with full images as a special case. This is most likely suboptimal. In this paper we focus on feature extraction and description over free-form regions and study the relationship with their fixed-form counterparts. Our main contributions are novel pooling techniques that capture the second-order statistics of local descriptors inside such free-form regions. We introduce second-order generalizations of average and max-pooling that together with appropriate non-linearities, derived from the mathematical structure of their embedding space, lead to state-of-the-art recognition performance in semantic segmentation experiments without any type of local feature coding. In contrast, we show that codebook-based local feature coding is more important when feature extraction is constrained to operate over regions that include both foreground and large portions of the background, as typical in image classification settings, whereas for high-accuracy localization setups, second-order pooling over free-form regions produces results superior to those of the winning systems in the contemporary semantic segmentation challenges, with models that are much faster in both training and testing.

Beyond the shortest path: Unsupervised domain adaptation by Sampling Subspaces along the Spline Flow

Recently, a particular paradigm [18] in the domain adaptation field has received considerable attention by introducing novel and important insights to the problem. In this case, the source and target domains are represented in the form of subspaces, which are treated as points on the Grassmann manifold. The geodesic curve between them is sampled to obtain intermediate points. Then a classifier is learnt using the projections of the data onto these subspaces. Despite its relevance and popularity, this paradigm [18] contains some limitations. Firstly, in real-world applications, that simple curve (i.e. shortest path) does not provide the necessary flexibility to model the domain shift between the training and testing data sets. Secondly, by using the geodesic curve, we are restricted to only one source domain, which does not allow to take fully advantage of the multiple datasets that are available nowadays. It is then, natural to ask whether this popular concept could be extended to deal with more complex curves and to integrate multi-sources domains. This is a hard problem considering the Riemannian structure of the space, but we propose a mathematically well-founded idea that enables us to solve it. We exploit the geometric insight of rolling maps [30] to compute a spline curve on the Grassmann manifold. The benefits of the proposed idea are demonstrated through several empirical studies on standard datasets. This novel concept allows to explicitly integrate multi-source domains while the previous one [18] uses the mean of all sources. This enables to model better the domain shift and take fully advantage of the training datasets.

Discriminative bayesian active shape models

This work presents a simple and very efficient solution to align facial parts in unseen images. Our solution relies on a Point Distribution Model (PDM) face model and a set of discriminant local detectors, one for each facial landmark. The patch responses can be embedded into a Bayesian inference problem, where the posterior distribution of the global warp is inferred in a

Rolling Riemannian Manifolds to Solve the Multi-class Classification Problem

#305;maximum a posteriori (MAP) sense. However, previous formulations do not model explicitly the covariance of the latent variables, which represents the confidence in the current solution. In our Discriminative Bayesian Active Shape Model (DBASM) formulation, the MAP global alignment is inferred by a Linear Dynamical System (LDS) that takes this information into account. The Bayesian paradigm provides an effective fitting strategy, since it combines in the same framework both the shape prior and multiple sets of patch alignment classifiers to further improve the accuracy. Extensive evaluations were performed on several datasets including the challenging Labeled Faces in the Wild (LFW). Face parts descriptors were also evaluated, including the recently proposed Minimum Output Sum of Squared Error (MOSSE) filter. The proposed Bayesian optimization strategy improves on the state-of-the-art while using the same local detectors. We also show that MOSSE filters further improve on these results.

Active Stereo Tracking of

In the past few years there has been a growing interest on geometric frameworks to learn supervised classification models on Riemannian manifolds [32, 28]. A popular framework, valid over any Riemannian manifold, was proposed in [32] for binary classification. Once moving from binary to multi-class classification this paradigm is not valid anymore, due to the spread of multiple positive classes on the manifold [28]. It is then natural to ask whether the multi-class paradigm could be extended to operate on a large class of Riemannian manifolds. We propose a mathematically well-founded classification paradigm that allows to extend the work in [32] to multi-class models, taking into account the structure of the space. The idea is to project all the data from the manifold onto an affine tangent space at a particular point. To mitigate the distortion induced by local diffeomorphisms, we introduce for the first time in the computer vision community a well-founded mathematical concept, so-called Rolling map [22, 17]. The novelty in this alternate school of thought is that the manifold will be firstly rolled (without slipping or twisting) as a rigid body, then the given data is unwrapped onto the affine tangent space, where the classification is performed.