James Tin-Yau Kwok

Domain Adaptation via Transfer Component Analysis

Domain adaptation allows knowledge from a source domain to be transferred to a different but related target domain. Intuitively, discovering a good feature representation across domains is crucial. In this paper, we first propose to find such a representation through a new learning method, transfer component analysis (TCA), for domain adaptation. TCA tries to learn some transfer components across domains in a reproducing kernel Hilbert space using maximum mean miscrepancy. In the subspace spanned by these transfer components, data properties are preserved and data distributions in different domains are close to each other. As a result, with the new representations in this subspace, we can apply standard machine learning methods to train classifiers or regression models in the source domain for use in the target domain. Furthermore, in order to uncover the knowledge hidden in the relations between the data labels from the source and target domains, we extend TCA in a semisupervised learning setting, which encodes label information into transfer components learning. We call this extension semisupervised TCA. The main contribution of our work is that we propose a novel dimensionality reduction framework for reducing the distance between domains in a latent space for domain adaptation. We propose both unsupervised and semisupervised feature extraction approaches, which can dramatically reduce the distance between domain distributions by projecting data onto the learned transfer components. Finally, our approach can handle large datasets and naturally lead to out-of-sample generalization. The effectiveness and efficiency of our approach are verified by experiments on five toy datasets and two real-world applications: cross-domain indoor WiFi localization and cross-domain text classification.

The pre-image problem in kernel methods

In this paper, we address the problem of finding the pre-image of a feature vector in the feature space induced by a kernel. This is of central importance in some kernel applications, such as on using kernel principal component analysis (PCA) for image denoising. Unlike the traditional method in which relies on nonlinear optimization, our proposed method directly finds the location of the pre-image based on distance constraints in the feature space. It is noniterative, involves only linear algebra and does not suffer from numerical instability or local minimum problems. Evaluations on performing kernel PCA and kernel clustering on the USPS data set show much improved performance.

Combination of images with diverse focuses using the spatial frequency

Abstract Image fusion attempts to combine complementary information from multiple images of the same scene, so that the resultant image is more suitable for human visual perception and computer-processing tasks such as segmentation, feature extraction and object recognition. This paper presents an approach that fuses images with diverse focuses by first decomposing the source images into blocks and then combining them by the use of spatial frequency. The algorithm is computationally simple and can be implemented in real-time applications. Experimental results show that the proposed method is superior to wavelet transform based methods in both objective and visual evaluations.

Using the discrete wavelet frame transform to merge Landsat TM and SPOT panchromatic images

In this paper, we propose a pixel level image fusion algorithm for merging Landsat thematic mapper (TM) images and SPOT panchromatic images. The two source images are first decomposed using the discrete wavelet frame transform (DWFT), which is both aliasing free and translation invariant. Wavelet coefficients from TM’s approximation subband and SPOT’s detail subbands are then combined together, and the fused image is reconstructed by performing the inverse DWFT. Experimental results show that the proposed approach outperforms methods based on the intensity-hue-saturation transform, principal component analysis and discrete wavelet transform in preserving spectral and spatial information, especially in situations where the source images are not perfectly registered. 2002 Elsevier Science B.V. All rights reserved.

Multifocus image fusion using artificial neural networks

Optical lenses, particularly those with long focal lengths, suffer from the problem of limited depth of field. Consequently, it is often difficult to obtain good focus for all objects in the picture. One possible solution is to take several pictures with different focus points, and then combine them together to form a single image. This paper describes an application of artificial neural networks to this pixel level multifocus image fusion problem based on the use of image blocks. Experimental results show that the proposed method outperforms the discrete wavelet transform based approach, particularly when there is a movement in the objects or misregistration of the source images.

Improved Nyström low-rank approximation and error analysis

Low-rank matrix approximation is an effective tool in alleviating the memory and computational burdens of kernel methods and sampling, as the mainstream of such algorithms, has drawn considerable attention in both theory and practice. This paper presents detailed studies on the Nyström sampling scheme and in particular, an error analysis that directly relates the Nyström approximation quality with the encoding powers of the landmark points in summarizing the data. The resultant error bound suggests a simple and efficient sampling scheme, the k-means clustering algorithm, for Nyström low-rank approximation. We compare it with state-of-the-art approaches that range from greedy schemes to probabilistic sampling. Our algorithm achieves significant performance gains in a number of supervised/unsupervised learning tasks including kernel PCA and least squares SVM.

Texture classification using the support vector machines

In recent years, support vector machines (SVMs) have demonstrated excellent performance in a variety of pattern recognition problems. In this paper, we apply SVMs for texture classification, using translation-invariant features generated from the discrete wavelet frame transform. To alleviate the problem of selecting the right kernel parameter in the SVM, we use a fusion scheme based on multiple SVMs, each with a different setting of the kernel parameter. Compared to the traditional Bayes classifier and the learning vector quantization algorithm, SVMs, and, in particular, the fused output from multiple SVMs, produce more accurate classification results on the Brodatz texture album.

A novel incremental principal component analysis and its application for face recognition

Principal component analysis (PCA) has been proven to be an efficient method in pattern recognition and image analysis. Recently, PCA has been extensively employed for face-recognition algorithms, such as eigenface and fisherface. The encouraging results have been reported and discussed in the literature. Many PCA-based face-recognition systems have also been developed in the last decade. However, existing PCA-based face-recognition systems are hard to scale up because of the computational cost and memory-requirement burden. To overcome this limitation, an incremental approach is usually adopted. Incremental PCA (IPCA) methods have been studied for many years in the machine-learning community. The major limitation of existing IPCA methods is that there is no guarantee on the approximation error. In view of this limitation, this paper proposes a new IPCA method based on the idea of a singular value decomposition (SVD) updating algorithm, namely an SVD updating-based IPCA (SVDU-IPCA) algorithm. In the proposed SVDU-IPCA algorithm, we have mathematically proved that the approximation error is bounded. A complexity analysis on the proposed method is also presented. Another characteristic of the proposed SVDU-IPCA algorithm is that it can be easily extended to a kernel version. The proposed method has been evaluated using available public databases, namely FERET, AR, and Yale B, and applied to existing face-recognition algorithms. Experimental results show that the difference of the average recognition accuracy between the proposed incremental method and the batch-mode method is less than 1%. This implies that the proposed SVDU-IPCA method gives a close approximation to the batch-mode PCA method

Moderating the outputs of support vector machine classifiers

In this paper, we extend the use of moderated outputs to the support vector machine (SVM) by making use of a relationship between SVM and the evidence framework. The moderated output is more in line with the Bayesian idea that the posterior weight distribution should be taken into account upon prediction, and it also alleviates the usual tendency of assigning overly high confidence to the estimated class memberships of the test patterns. Moreover, the moderated output derived here can be taken as an approximation to the posterior class probability. Hence, meaningful rejection thresholds can be assigned and outputs from several networks can be directly compared. Experimental results on both artificial and real-world data are also discussed.

The evidence framework applied to support vector machines

In this paper, we show that training of the support vector machine (SVM) can be interpreted as performing the level 1 inference of MacKays evidence framework.We further on show that levels 2 and 3 of the evidence framework can also be applied to SVMs. This integration allows automatic adjustment of the regularization parameter and the kernel parameter to their near-optimal values. Moreover, it opens up a wealth of Bayesian tools for use with SVMs. Performance of this method is evaluated on both synthetic and real-world data sets.