Rongguo Zhang

Multiple Instance Learning Based Method for Similar Handwritten Chinese Characters Discrimination

This paper proposes a Multiple Instance Learning based method for similar handwritten Chinese characters discrimination. The similar handwritten Chinese characters recognition problem is first defined as a Multiple-instance learning problem. Then the problem is solved by the AdaBoost framework. The proposed method selects some self-adapting critical regions as weak classifiers, and therefore it is more suitable for the wide variability of writing styles. Our experimental results demonstrate that the proposed method outperforms the other state-of-the-art methods.

A strategy of classification via sparse dictionary learned by non-negative K-SVD

In recent years there is a growing interest in the study of sparse representation for signals. This article extends this research into a novel model for object classification tasks. In this model, we first apply the non-negative K-SVD algorithm to learning the discriminative dictionaries using very few training samples and then represent a test image as a linear combination of atoms from these learned dictionaries based on the non-negative variation of Basis Pursuit (BP). Finally, we achieve the classification purpose by analyzing the sparse weighting coefficients. Our strategy of classification is very simple and does not ask much for the training samples. Our model is tested on two benchmark data sets Caltech-101 and UIUC-car. In both datasets, Our approach achieves the comparable performance. The idea in this paper strengthens the case for using this model in computer vision further.

Modified Two-Class LDA Based Compound Distance for Similar Handwritten Chinese Characters Discrimination

This paper proposes a modified two-class LDA based compound distance for similar handwritten Chinese characters discrimination. First the definition of the Intersecting Subspace (IS) between two classes and the modified between-class scatter matrix is given. Then we prove that the modified between-class scatter matrix can supply additional information. Our experiments demonstrate that the additional information can be used to discriminate points in the IS and the proposed method outperforms the previous LDA based method.

A local learning based Image-To-Class distance for image classification

Image-To-Class distance is first proposed in Naive-Bayes Nearest-Neighbor. NBNN is a feature-based image classifier, and can achieve impressive classification accuracy. However, the performance of NBNN relies heavily on the large number of training samples. If using small number of training samples, the performance will degrade. The goal of this paper is to address this issue. The main contribution of this paper is that we propose a robust Image-to-Class distance by local learning. We define the patch-to-class distance as the distance between the input patch to its nearest neighbor in one class, which is reconstructed in the local manifold space; and then our image-to-class distance is the sum of patch-to-class distance. Furthermore, we take advantage of large-margin metric learning framework to obtain a proper Mahalanobis metric for each class. We evaluate the proposed method on four benchmark datasets: Caltech, Corel, Scene13, and Graz. The results show that our defined Image-To-Class Distance is more robust than NBNN and Optimal-NBNN, and by combining with the learned metric for each class, our method can achieve significant improvement over previous reported results on these datasets.

Quadratic-chi similarity metric learning for histogram feature

Histogram features, such as SIFT, HOG, LBP et al, are widely used in modern computer vision algorithms. According to [18], chi-square distance is an effective measure for comparing histogram features. In this paper, we propose a new method, named the Quadric-chi similarity metric learning (QCSML) for histogram features. The main contribution of this paper is that we propose a new metric learning method based on chi-square distance, in contrast with traditional Mahalanobis distance metric learning methods. The use of quadric-chi similarity in our method leads to an effective learning algorithm. Our method is tested on SIFT features for face identification, and compared with the state-of-art metric learning method (LDML) on the benchmark dataset, the Labeled Faces in the Wild (LFW). Experimental results show that our method can achieve clear performance gains over LDML.

Data Transformation of the Histogram Feature in Object Detection

Detecting objects in images is very important for several application domains in computer vision. This paper presents an experimental study on data transformation of the feature vector in object detection. We use the modified Pyramid of Histograms of Orientation Gradients descriptor and the SVM classifier to form an object detection model. We apply a simple transformation to the histogram features before training and testing. This transformation equals a small change in the kernel function for Support Vector Machines. This change is much quicker than the χ2 kernel, but obtains better results. Experimental evaluations on the UIUC Image Database and TU Darmstadt Database show that the transformed features perform better than the raw features, and this transformation improves the linear separability of the histogram feature.