Regularization plays an important role in machine learning systems. We propose a novel methodology for model regularization using random projection. We demonstrate the technique on neural networks, since such models usually comprise a very large number of parameters, calling for strong regularizers. It has been shown recently that neural networks are sensitive to two kinds of samples: (i) adversarial samples, which are generated by imperceptible perturbations of previously correctly-classified samples—yet the network will misclassify them; and (ii) fooling samples, which are completely unrecognizable, yet the network will classify them with extremely high confidence. In this paper, we show how robust neural networks can be trained using random projection. We show that while random projection acts as a strong regularizer, boosting model accuracy similar to other regularizers, such as weight decay and dropout, it is far more robust to adversarial noise and fooling samples. We further show that random projection also helps to improve the robustness of traditional classifiers, such as Random Forrest and Gradient Boosting Machines.

Many typical applications of object detection operate within a prescribed false-positive range. In this situation the performance of a detector should be assessed on the basis of the area under the ROC curve over that range, rather than over the full curve, as the performance outside the prescribed range is irrelevant. This measure is labelled as the partial area under the ROC curve (pAUC). We propose a novel ensemble learning method which achieves a maximal detection rate at a user-defined range of false positive rates by directly optimizing the partial AUC using structured learning. In addition, in order to achieve high object detection performance, we propose a new approach to extracting low-level visual features based on spatial pooling. Incorporating spatial pooling improves the translational invariance and thus the robustness of the detection process. Experimental results on both synthetic and real-world data sets demonstrate the effectiveness of our approach, and we show that it is possible to train state-of-the-art pedestrian detectors using the proposed structured ensemble learning method with spatially pooled features. The result is the current best reported performance on the Caltech-USA pedestrian detection dataset.

Traffic scene perception (TSP) aims to extract accurate real-time on-road environment information, which involves three phases: detection of objects of interest, recognition of detected objects, and tracking of objects in motion. Since recognition and tracking often rely on the results from detection, the ability to detect objects of interest effectively plays a crucial role in TSP. In this paper, we focus on three important classes of objects: traffic signs, cars, and cyclists. We propose to detect all the three important objects in a single learning-based detection framework. The proposed framework consists of a dense feature extractor and detectors of three important classes. Once the dense features have been extracted, these features are shared with all detectors. The advantage of using one common framework is that the detection speed is much faster, since all dense features need only to be evaluated once in the testing phase. In contrast, most previous works have designed specific detectors using different features for each of these three classes. To enhance the feature robustness to noises and image deformations, we introduce spatially pooled features as a part of aggregated channel features. In order to further improve the generalization performance, we propose an object subcategorization method as a means of capturing the intraclass variation of objects. We experimentally demonstrate the effectiveness and efficiency of the proposed framework in three detection applications: traffic sign detection, car detection, and cyclist detection. The proposed framework achieves the competitive performance with state-of-the-art approaches on several benchmark data sets.

We propose an effective structured learning based approach to the problem of person re-identification which outperforms the current state-of-the-art on most benchmark data sets evaluated. Our framework is built on the basis of multiple low-level hand-crafted and high-level visual features. We then formulate two optimization algorithms, which directly optimize evaluation measures commonly used in person re-identification, also known as the Cumulative Matching Characteristic (CMC) curve. Our new approach is practical to many real-world surveillance applications as the re-identification performance can be concentrated in the range of most practical importance. The combination of these factors leads to a person re-identification system which outperforms most existing algorithms. More importantly, we advance state-of-the-art results on person re-identification by improving the rank-1 recognition rates from 40% to 50% on the iLIDS benchmark, 16% to 18% on the PRID2011 benchmark, 43% to 46% on the VIPeR benchmark, 34% to 53% on the CUHK01 benchmark and 21% to 62% on the CUHK03 benchmark.

The use of high-dimensional features has become a normal practice in many computer vision applications. The large dimension of these features is a limiting factor upon the number of data points, which may be effectively stored and processed, however. We address this problem by developing a novel approach to learning a compact binary encoding, which exploits both pairwise proximity and class-label information on training data set. Exploiting this extra information apairwise proximity and class-label information on training data set. Exploiting this extra information allows the development of encodings which, although compact, outperform the original high-dimensional features in terms of final classification or retrieval performance. The method is general, in that it is applicable to both nonparametric and parametric learning methods. This generality means that the embedded features are suitable for a wide variety of computer vision tasks, such as image classification allows the development of encodings which, although compact, outperform the original high-dimensional features in terms of final classification or retrieval performance. The method is general, in that it is applicable to both nonparametric and parametric learning methods. This generality means that the embedded features are suitable for a wide variety of computer vision tasks, such as image classification and content-based image retrieval. Experimental results demonstrate that the new compact descriptor achieves an accuracy comparable to, and in some cases better than, the visual descriptor in the original space despite being significantly more compact. Moreover, any convex loss function and convex regularization penalty (e.g., lpnorm with p ≥ 1) can be incorporated into the framework, which provides future flexibility.

Cascade classifiers are one of the most important contributions to real-time object detection. Nonetheless, there are many challenging problems arising in training cascade detectors. One common issue is that the node classifier is trained with a symmetric classifier. Having a low misclassification error rate does not guarantee an optimal node learning goal in cascade classifiers, i.e., an extremely high detection rate with a moderate false positive rate. In this work, we present a new approach to train an effective node classifier in a cascade detector. The algorithm is based on two key observations: 1) Redundant weak classifiers can be safely discarded; 2) The final detector should satisfy the asymmetric learning objective of the cascade architecture. To achieve this, we separate the classifier training into two steps: finding a pool of discriminative weak classifiers/features and training the final classifier by pruning weak classifiers which contribute little to the asymmetric learning criterion (asymmetric classifier construction). Our model reduction approach helps accelerate the learning time while achieving the pre-determined learning objective. Experimental results on both face and car data sets verify the effectiveness of the proposed algorithm. On the FDDB face data sets, our approach achieves the state-of-the-art performance, which demonstrates the advantage of our approach.

We propose a novel boosting approach to multiclass classification problems, in which multiple classes are distinguished by a set of random projection matrices in essence. The approach uses random projections to alleviate the proliferation of binary classifiers typically required to perform multiclass classification. The result is a multiclass classifier with a single vector-valued parameter, irrespective of the number of classes involved. Two variants of this approach are proposed. The first method randomly projects the original data into new spaces, while the second method randomly projects the outputs of learned weak classifiers. These methods are not only conceptually simple but also effective and easy to implement. A series of experiments on synthetic, machine learning, and visual recognition data sets demonstrate that our proposed methods could be compared favorably with existing multiclass boosting algorithms in terms of both the convergence rate and classification accuracy.

Many typical applications of object detection operate within a prescribed false-positive range. In this situation the performance of a detector should be assessed on the basis of the area under the ROC curve over that range, rather than over the full curve, as the performance outside the range is irrelevant. This measure is labelled as the partial area under the ROC curve (pAUC). Effective cascade-based classification, for example, depends on training node classifiers that achieve the maximal detection rate at a moderate false positive rate, e.g., around 40% to 50%. We propose a novel ensemble learning method which achieves a maximal detection rate at a user-defined range of false positive rates by directly optimizing the partial AUC using structured learning. By optimizing for different ranges of false positive rates, the proposed method can be used to train either a single strong classifier or a node classifier forming part of a cascade classifier. Experimental results on both synthetic and real-world data sets demonstrate the effectiveness of our approach, and we show that it is possible to train state-of-the-art pedestrian detectors using the proposed structured ensemble learning method.

We present a novel formulation of fully corrective boosting for multi-class classification problems with the awareness of sharing features. Our multi-class boosting is solved in a single optimization problem. In order to share features across different classes, we introduce the mixed-norm regularization, which promotes group sparsity, into boosting. We then derive the Lagrange dual problems which enable us to design fully corrective multi-class algorithms using the primal-dual optimization technique. We show that sharing features across classes can improve classification performance and efficiency. We empirically show that in many cases, the proposed multi-class boosting generalizes better than a range of competing multi-class boosting algorithms due to the capability of feature sharing. Experimental results on machine learning data, visual scene and object recognition demonstrate the efficiency and effectiveness of proposed algorithms and validate our theoretical findings.

The ability to efficiently and accurately detect objects plays a very crucial role for many computer vision tasks. Recently, offline object detectors have shown a tremendous success. However, one major drawback of offline techniques is that a complete set of training data has to be collected beforehand. In addition, once learned, an offline detector cannot make use of newly arriving data. To alleviate these drawbacks, online learning has been adopted with the following objectives: 1) the technique should be computationally and storage efficient; 2) the updated classifier must maintain its high classification accuracy. In this paper, we propose an effective and efficient framework for learning an adaptive online greedy sparse linear discriminant analysis model. Unlike many existing online boosting detectors, which usually apply exponential or logistic loss, our online algorithm makes use of linear discriminant analysis' learning criterion that not only aims to maximize the class-separation criterion but also incorporates the asymmetrical property of training data distributions. We provide a better alternative for online boosting algorithms in the context of training a visual object detector. We demonstrate the robustness and efficiency of our methods on handwritten digit and face data sets. Our results confirm that object detection tasks benefit significantly when trained in an online manner.

Real-time object detection has many computer vision applications. Since Viola and Jones proposed the first real-time AdaBoost based face detection system, much effort has been spent on improving the boosting method. In this work, we first show that feature selection methods other than boosting can also be used for training an efficient object detector. In particular, we introduce greedy sparse linear discriminant analysis (GSLDA) for its conceptual simplicity and computational efficiency; and slightly better detection performance is achieved compared with . Moreover, we propose a new technique, termed boosted greedy sparse linear discriminant analysis (BGSLDA), to efficiently train a detection cascade. BGSLDA exploits the sample reweighting property of boosting and the class-separability criterion of GSLDA. Experiments in the domain of highly skewed data distributions (e.g., face detection) demonstrate that classifiers trained with the proposed BGSLDA outperforms AdaBoost and its variants. This finding provides a significant opportunity to argue that AdaBoost and similar approaches are not the only methods that can achieve high detection results for real-time object detection.

Face detection plays an important role in many vision applications. Since Viola and Jones [1] proposed the first real-time AdaBoost based object detection system, much effort has been spent on improving the boosting method. In this work, we first show that feature selection methods other than boosting can also be used for training an efficient object detector. In particular, we have adopted Greedy Sparse Linear Discriminant Analysis (GSLDA) [2] for its computational efficiency; and slightly better detection performance is achieved compared with [1]. Moreover, we propose a new technique, termed Boosted Greedy Sparse Linear Discriminant Analysis (BGSLDA), to efficiently train object detectors. BGSLDA exploits the sample re-weighting property of boosting and the class-separability criterion of GSLDA. Experiments in the domain of highly skewed data distributions, e.g., face detection, demonstrates that classifiers trained with the proposed BGSLDA outperforms AdaBoost and its variants. This finding provides a significant opportunity to argue that Adaboost and similar approaches are not the only methods that can achieve high classification results for high dimensional data such as object detection.

Efficiently and accurately detecting pedestrians plays a crucial role in many vision applications such as video surveillance, multimedia retrieval and smart car etc. In order to find the right feature for this task, we first present a comprehensive experimental study on pedestrian detection using state-of-the-art locally-extracted features. Building upon our findings, we propose a new, simpler pedestrian detecting framework based on the covariance features. We conduct feature selection and weak classifier training in the Euclidean space for faster computation. To this end, two machine learning algorithms have been designed: AdaBoost with weighted Fisher linear discriminant analysis (WLDA) based weak classifiers and Greedy Sparse Linear Discriminant Analysis (GSLDA). To further accelerate the detection, we employ a faster strategy, multiple cascade layers with heterogeneous features, to exploit the efficiency of the Haar-like features and the discriminative power of the covariance features. Experimental results shown on different datasets prove that the new pedestrian detection is not only comparable to the performance of the state-of-the-art pedestrian detectors but it also performs at a faster speed.

Face detection in images is very important for many multimedia applications. Haar-like wavelet features have become dominant in face detection because of their tremendous success since Viola and Jones [1] proposed their AdaBoost based detection system. While Haar features' simplicity makes rapid computation possible, its discriminative power is limited. As a consequence, a large training dataset is required to train a classifier. This may hamper its application in scenarios that a large labeled dataset is difficult to obtain. In this work, we address the problem of learning to detect faces from a small set of training examples. In particular, we propose to use co- variance features. Also for better classification performance, linear hyperplane classifier based on Fisher discriminant analysis (FDA) is proffered. Compared with the decision stump, FDA is more discriminative and therefore fewer weak learners are needed. We show that the detection rate can be significantly improved with covariance features on a small dataset (a few hundred positive examples), compared to Haar features used in current most face detection systems.

This paper presents an experimental study on pedestrian detection using state-of-the-art local feature extraction and support vector machine (SVM) classifiers. The performance of pedestrian detection using region covariance, histogram of oriented gradients (HOG) and local receptive fields (LRF) feature descriptors is experimentally evaluated. The experiments are performed on both the benchmarking dataset used in [1] and the MIT CBCL dataset. Both can be publicly accessed. The experimental results show that region covariance features with radial basis function (RBF) kernel SVM and HOG features with quadratic kernel SVM outperform the combination of LRF features with quadratic kernel SVM reported in [1].

Efficiently and accurately detecting pedestrians plays a very important role in many computer vision applications such as video surveillance and smart cars. In order to find the right feature for this task, we first present a comprehensive experimental study on pedestrian detection using state-of-the-art locally extracted features (e.g., local receptive fields, histogram of oriented gradients, and region covariance). Building upon the findings of our experiments, we propose a new, simpler pedestrian detector using the covariance features. Unlike the work in [1], where the feature selection and weak classifier training are performed on the Riemannian manifold, we select features and train weak classifiers in the Euclidean space for faster computation. To this end, AdaBoost with weighted Fisher linear discriminant analysis-based weak classifiers are designed. A cascaded classifier structure is constructed for efficiency in the detection phase. Experiments on different datasets prove that the new pedestrian detector is not only comparable to the state-of-the-art pedestrian detectors but it also performs at a faster speed. To further accelerate the detection, we adopt a faster strategy-multiple layer boosting with heterogeneous features-to exploit the efficiency of the Haar feature and the discriminative power of the covariance feature. Experiments show that, by combining the Haar and covariance features, we speed up the original covariance feature detector [1] by up to an order of magnitude in detection time with a slight drop in detection performance.