This paper addresses the problem of learning from very large databases where batch learning is impractical or even infeasible. Bootstrap is a popular technique applicable in such situations. We show that sampling strategy used for bootstrapping has a significant impact on the resulting classifier performance. We design a new general sampling strategy ”quasi-random weighted sampling + trimming” (QWS+) that includes well established strategies as special cases. The QWS+ approach minimizes the variance of hypothesis error estimate and leads to significant improvement in performance compared to standard sampling techniques. The superior performance is demonstrated on several problems including profile and frontal face detection.

This work investigates the problem of robust, longterm visual tracking of unknown objects in unconstrained environments. It therefore must cope with frame-cuts, fast camera movements and partial/total object occlusions/dissapearances. We propose a new approach, called Tracking-Modeling-Detection (TMD) that closely integrates adaptive tracking with online learning of the object-specific detector. Starting from a single click in the first frame, TMD tracks the selected object by an adaptive tracker. The trajectory is observed by two processes (growing and pruning event) that robustly model the appearance and build an object detector on the fly. Both events make errors, the stability of the system is achieved by their cancelation. The learnt detector enables re-initialization of the tracker whenever previously observed appearance reoccurs. We show the real-time learning and classification is achievable with random forests. The performance and the long-term stability of TMD is demonstrated and evaluated on a set of challenging video sequences with various objects such as cars, people and animals. 1.

Boosting has been widely applied in computer vision, especially after Viola and Jones’s seminal work [23]. The marriage of rectangular features and integral-imageenabled fast computation makes boosting attractive for many vision applications. However, this popular way of applying boosting normally employs an exhaustive feature selection scheme from a very large hypothesis pool, which results in a less-efficient learning process. Furthermore, this poses additional constraint on applying boosting in an online fashion, where feature re-selection is often necessary because of varying data characteristic, but yet impractical due to the huge hypothesis pool. This paper proposes a gradient-based feature selection approach. Assuming a generally trained feature set and labeled samples are given, our approach iteratively updates each feature using the gradient descent, by minimizing the weighted least square error between the estimated feature response and the true label. In addition, we integrate the gradient-based feature selection with an online boosting framework. This new online boosting algorithm not only provides an efficient way of updating the discriminative feature set, but also presents a unified objective for both feature selection and weak classifier updating. Experiments on the person detection and tracking applications demonstrate the effectiveness of our proposal. 1.

Random Forests (RFs) are frequently used in many computer vision and machine learning applications. Their popularity is mainly driven by their high computational efficiency during both training and evaluation while still achieving state-of-the-art results. However, in most applications RFs are used off-line. This limits their usability for many practical problems, for instance, when training data arrives sequentially or the underlying distribution is continuously changing. In this paper, we propose a novel on-line random forest algorithm. We combine ideas from on-line bagging, extremely randomized forests and propose an on-line decision tree growing procedure. Additionally, we add a temporal weighting scheme for adaptively discarding some trees based on their out-of-bag-error in given time intervals and consequently growing of new trees. The experiments on common machine learning data sets show that our algorithm converges to the performance of the off-line RF. Additionally, we conduct experiments for visual tracking, where we demonstrate real-time state-of-the-art performance on wellknown scenarios and show good performance in case of occlusions and appearance changes where we outperform trackers based on on-line boosting. Finally, we demonstrate the usability of on-line RFs on the task of interactive realtime segmentation. 1.

In this paper we present a simple approach for person detection in surveillance for static cameras. The basic idea is to train a separate classifier for each image location which has only to discriminate the object from the background at a specific location. This is a considerably simpler problem than the detection of persons on arbitrary backgrounds. Therefore, we use adaptive classifiers which are trained online. Due to the reduced complexity we can use a simple update strategy that requires only a few positive samples and is stable by design. This is an essential property for real world applications which require operation for 24 hours a day, 7 days a week. We demonstrate and evaluate the method on publicly available sequences and compare it to state-of-theart methods which reveals that despite the simple strategy the obtained performance is competitive.

In this paper we present an adaptive but robust object detector for static cameras by introducing classifier grids. Instead of using a sliding window for object detection we propose to train a separate classifier for each image location, obtaining a very specific object detector with a low false alarm rate. For each classifier corresponding to a grid element we estimate two generative representations in parallel, one describing the object’s class and one describing the background. These are combined in order to obtain a discriminative model. To enable to adapt to changing environments these classifiers are learned on-line (i.e., boosting). Continuously learning (24 hours a day, 7 days a week) requires a stable system. In our method this is ensured by a fixed object representation while updating only the representation of the background. We demonstrate the stability in a long-term experiment by running the system for a whole week, which shows a stable performance over time. In addition, we compare the proposed approach to state-of-the-art methods in the field of person and car detection. In both cases we obtain competitive results. 1.

Abstract. The sparse representation has been widely used in many areas and utilized for visual tracking. Tracking with sparse representation is formulated as searching for samples with minimal reconstruction errors from learned template subspace. However, the computational cost makes it unsuitable to utilize high dimensional advanced features which are often important for robust tracking under dynamic environment. Based on the observations that a target can be reconstructed from several templates, and only some of the features with discriminative power are significant to separate the target from the background, we propose a novel online tracking algorithm with two stage sparse optimization to jointly minimize the target reconstruction error and maximize the discriminative power. As the target template and discriminative features usually have temporal and spatial relationship, dynamic group sparsity (DGS) is utilized in our algorithm. The proposed method is compared with three state-of-art trackers using five public challenging sequences, which exhibit appearance changes, heavy occlusions, and pose variations. Our algorithm is shown to outperform these methods. 1

This paper proposes an algorithm for online feature selection which improves robustness to occlusions by referring to a localized generative appearance model. Discriminative classifiers based on feature extraction have classically either prepared a fixed prior model by training offline, or continually adapted their classification parameters to any apparent appearance changes. By combining the attractive qualities of each approach, our framework can cope with appearance changes of a target object and will maintain proximity to a static appearance model. Our main contribution is the use of a generative model to guide the online feature selection to regions of an image which maintain a valid appearance. The generative model exhibits the properties of non-negativity, localization and orthogonality. We demonstrate the system in a tracking framework to show improved tracking performance through occlusions. 1

In recent years, boosting has been successfully applied to many practical problems in pattern recognition and computer vision fields such as object detection and tracking. As boosting is an offline training process with beforehand collected data, once learned, it cannot make use of any newly arriving ones. However, an offline boosted detector is to be exploited online and inevitably there must be some special cases that are not covered by those beforehand collected training data. As a result, the inadaptable detector often performs badly in diverse and changeful environments which are ordinary for many real-life applications. To alleviate this problem, this paper proposes an incremental learning algorithm to effectively adjust a boosted strong classifier with domain-partitioning weak hypotheses to online samples, which adopts a novel

This paper addresses the problem of simultaneous tracking of multiple targets in a video. We first apply object detectors to every video frame. Pairs of detection responses from every two consecutive frames are then used to build a graph of tracklets. The graph helps transitively link the best matching tracklets that do not violate hard and soft contextual constraints between the resulting tracks. We prove that this data association problem can be formulated as finding the maximum-weight independent set (MWIS) of the graph. We present a new, polynomial-time MWIS algorithm, and prove that it converges to an optimum. Similarity and contextual constraints between object detections, used for data association, are learned online from object appearance and motion properties. Long-term occlusions are addressed by iteratively repeating MWIS to hierarchically merge smaller tracks into longer ones. Our results demonstrate advantages of simultaneously accounting for soft and hard contextual constraints in multitarget tracking. We outperform the state of the art on the benchmark datasets. 1.