Susumu Kubota

A Global Optimization Algorithm for Real-Time On-Board Stereo Obstacle Detection Systems

A fast and robust stereo algorithm for on-board obstacle detection systems is proposed. The proposed method finds the optimum road-obstacle boundary which provides the most consistent interpretation of the input stereo image pair. Global optimization combined with a robust matching measure enables stable detection of obstacles under various circumstances, such as heavy rain and severe lighting conditions. The processing time for VGA size image pair is about 15 msec on a 3.6 GHz pentium IV processor, which is fast enough for realtime applications.

Object classification using heterogeneous co-occurrence features

Co-occurrence features are effective for object classification because observing co-occurrence of two events is far more informative than observing occurrence of each event separately. For example, a color co-occurrence histogram captures co-occurrence of pairs of colors at a given distance while a color histogram just expresses frequency of each color. As one of such co-occurrence features, CoHOG (co-occurrence histograms of oriented gradients) has been proposed and a method using CoHOG with a linear classifier has shown a comparable performance with state-of-the-art pedestrian detection methods. According to recent studies, it has been suggested that combining heterogeneous features such as texture, shape, and color is useful for object classification. Therefore, we introduce three heterogeneous features based on co-occurrence called color-CoHOG, CoHED, and CoHD, respectively. Each heterogeneous features are evaluated on the INRIA person dataset and the Oxford 17/102 category flower datasets. The experimental results show that color-CoHOG is effective for the INRIA person dataset and CoHED is effective for the Oxford flower datasets. By combining above heterogeneous features, the proposed method achieves comparable classification performance to state-of-the-art methods on the above datasets. The results suggest that the proposed method using heterogeneous features can be used as an off-the-shelf method for various object classification tasks.

Random ensemble metrics for object recognition

This paper presents a novel and generic approach for metric learning, random ensemble metrics (REMetric). To improve generalization performance, we introduce the concept of ensemble learning to the metric learning scheme. Unlike previous methods, our method does not optimize the global objective function for the whole training data. It learns multiple discriminative projection vectors obtained from linear support vector machines (SVM) using randomly subsampled training data. The final metric matrix is then obtained by integrating these vectors. As a result of using SVM, the learned metric has an excellent scalability for the dimensionality of features. Therefore, it does not require any prior dimensionality reduction techniques such as PCA. Moreover, our method allows us to unify dimensionality reduction and metric learning by controlling the number of the projection vectors. We demonstrate through experiments, that our method can avoid overfitting even though a relatively small number of training data is provided. The experiments are performed with three different datasets; the Viewpoint Invariant Pedestrian Recognition (VIPeR) dataset, the Labeled Face in the Wild (LFW) dataset and the Oxford 102 category flower dataset. The results show that our method achieves equivalent or superior performance compared to existing state-of-the-art metric learning methods.

Cat face detection with two heterogeneous features

In this paper, we propose a generic and efficient object detection framework based on two heterogeneous features and demonstrate effectiveness of our method for a cat face detection problem. Simple Haar-like features with AdaBoost are fast to compute but they are not discriminative enough to deal with complicated shape and texture. Therefore, we cascade joint Haar-like features with AdaBoost and CoHOG descriptors with a linear classifier. Since the CoHOG descriptors are extremely high dimensional pattern descriptors based on gradient orientations, they have a strong classification capability to represent various cat face patterns. The combination of these two distinct classifiers enables fast and accurate cat face detection. The experimental result with about 10,000 cat images shows that our method gives better performance in comparison with the state-of-the-art cat head detection method, although our method does not exploit any cat specific characteristics.

Uncertainty model of the gradient constraint and quantitative reliability measures of optical flow

When computing optical flow based on a gradient method, the quantitative uncertainty model of the gradient constraint is indispensable for deriving quantitative reliability measures of optical flow. Although the gradient constraint has been considered to be derived from the Taylor expansion of image intensity, such an interpretation has several drawbacks in the estimation of uncertainty. In this paper, another interpretation of the gradient constraint is presented based on least squares approximation, and a quantitative uncertainty model is derived from it. Quantitative estimation of the uncertainty of the gradient constraint enables accurate computation of optical flow and provides quantitative reliability measures of optical flow. The validity of the proposed model and the quantitative reliability measures are verified through experiments.

Large Margin Discriminant Hashing for Fast k-Nearest Neighbor Classification

Since the k-nearest neighbor (k-NN) classification is computationally demanding in terms of time and memory, approximate nearest neighbor (ANN) algorithms that utilize dimensionality reduction and hashing are gathering interest. Dimensionality reduction saves memory usage for storing training patterns and hashing techniques significantly reduce the computation required for distance calculation. Several ANN methods have been proposed which make k-NN classification applicable to those tasks that have a large number of training patterns with very high-dimensional feature. Though conventional ANN methods try to approximate Euclidean distance calculation in the original high-dimensional feature space with much lower-dimensional subspace, the Euclidean distance in the original feature space is not necessarily optimal for classification. According to the recent studies, metric learning is effective to improve accuracy of the k-NN classification. In this paper, Large Margin Discriminative Hashing (LMDH) method, which projects input patterns into low dimensional subspace with the optimized metric for the k-NN classification, is proposed.

A discriminative learning criterion for the overall optimization of error and reject

The minimum classification error (MCE) criterion has been commonly used for discriminative learning but there is intrinsic difficulty in applying it to gradient descent methods. As the complete description of classification performance is given by the error-reject tradeoff, we augment the MCE criterion not only to include but also reject errors and show that it leads to a smooth loss function which is suitable for gradient descent methods. The proposed criterion provides a quantitative justification for the loss function in terms of the classification performance. The loss function is adaptively optimized based on the empirical distribution of the classifier output at each iteration of the learning procedure. Since the proposed method does not need any manual parameter tuning, it is exempt from time consuming trial and error. Nevertheless, experimental results show that the results of the proposed method are better than those of the MCE method with the best tuned parameters. A comparison with the maximum mutual information criterion shows that the proposed criterion has better outlier resistance than that of the MMI.

Fast and Memory Efficient Online Handwritten Strokes Retrieval Using Binary Descriptor

We propose a fast, memory-efficient online handwriting search method that uses handwritten strokes as a query and finds matches from among handwritten documents. The proposed method is language-independent, so not only words but also figures and symbols can be queried. We introduce a compact binary descriptor to lower computational resource load. A metric learning method enables derivation of a discriminative binary descriptor from directional densities of handwritten strokes. Experiments indicate that the proposed method is faster, more memory efficient, and exhibits more accurate search performance than a conventional method that employs directional densities. For 200 handwritten documents, the proposed method completed query searches within 1 s using a 1.3 GHz Tegra 3 CPU.