Labelling 2-D geometric primitives using probabilistic relaxation:reducing the computational requirements

In previous work, an algorithm for matching geometric features was developed. Although the method worked well, it was demanding in computational resources when applied to large problems. The authors now present methods for reducing the computational requirements, without significantly affecting the reliability of the algorithm     

Sum versus vote fusion in multiple classifier systems

Amidst the conflicting experimental evidence of superiority of one over the other, we investigate the Sum and majority Vote combining rules in a two class case, under the assumption of experts being of equal strength and estimation errors conditionally independent and identically distributed. We show, analytically, that, for Gaussian estimation error distributions, Sum always outperforms Vote. For heavy tail distributions, we demonstrate by simulation that Vote may outperform Sum. Results on synthetic data confirm the theoretical predictions. Experiments on real data support the general findings, but also show the effect of the usual assumptions of conditional independence, identical error distributions, and common target outputs of the experts not being fully satisfied.     

Classification of sets of mixed pixels with the hypothesis-testingHough transform

A new method is presented for mixed pixel classification where the classification of groups of mixed pixels is achieved by using the hypothesis-testing Hough transform. The motivation of the work is that some other estimation methods based on robust statistics, such as the standard Hough transform, have been criticised that, although they can cope with the presence of outliers, they give poor performance in the absence of outliers in comparison to the least-squares-error method. The method proposed in the paper is demonstrated using simulated data and proved to perform equally well in the presence and in the absence of outliers. It is also applied to real Landsat TM data     

Automatic annotation of tennis games: An integration of audio,vision, and learning

Fully automatic annotation of tennis game using broadcast video is a task with a great potential but with enormous challenges. In this paper we describe our approach to this task, which integrates computer vision, machine listening, and machine learning. At the low level processing, we improve upon our previously proposed state-of-the-art tennis ball tracking algorithm and employ audio signal processing techniques to detect key events and construct features for classifying the events. At high level analysis, we model event classification as a sequence labelling problem, and investigate four machine learning techniques using simulated event sequences. Finally, we evaluate our proposed approach on three real world tennis games, and discuss the interplay between audio, vision and learning. To the best of our knowledge, our system is the only one that can annotate tennis game at such a detailed level.     

Low-Temperature Drying Using a Versatile Heat Pump Dehumidifier

A drying system incorporating a commercial 2.3 kW heat pump dehumidifier was designed and constructed. The HPD was equipped with an external water-cooled condenser that rejected excess heat out of the system. The design of the system allowed for conducting drying with recirculation of air as well as use of electrical heaters. In an open mode, the drying could be carried out simultaneously with room dehumidification and water heating in the secondary condenser. Drying experiments were conducted with apple and comparisons were made between HPD assisted drying (partial and complete) and hot air drying (at 45 and 65°C). The HPD dried fruit exhibited better rehydration properties than the hot air dried samples. Water activity of the HPD dried samples was noticeably lower than that of the hot air dried samples at the same water content, indicating that the residual moisture was probably held under higher tension in the former. In terms of energy consumption, the process of HPD assisted drying is more expensive as much of the energy input is rejected at the secondary condenser as excess heat.     

Pose-invariant face recognition by matching on multi-resolution MRFs linked by supercoupling transform

Recognition of faces in arbitrary pose is addressed in this paper. For this task, an MRF-based classification approach is proposed which employs the energy of the established match between a pair of images as a criterion of goodness-of-match. By incorporating an image matching method as part of the recognition process, the system is made robust to moderate global spatial transformations. The approach draws on a method [1] which has the potential to cope with pose changes but a direct application of which suffers from several shortcomings. In order to overcome these problems, a number of enhancements are proposed. First, by adopting a multi-scale relaxation scheme based on super coupling transform, the inference using sequential tree re-weighted message passing approach [2] is accelerated. Next, by taking advantage of a statistical shape prior for the matching, the results are regularized and constrained, making the system robust to spurious structures and outliers. For classification, both textural and structural similarities of the facial images are taken into account. The method is evaluated on two databases and promising results are obtained.     

Incremental Linear Discriminant Analysis Using Sufficient Spanning Sets and Its Applications

This paper presents an incremental learning solution for Linear Discriminant Analysis (LDA) and its applications to object recognition problems. We apply the sufficient spanning set approximation in three steps i.e. update for the total scatter matrix, between-class scatter matrix and the projected data matrix, which leads an online solution which closely agrees with the batch solution in accuracy while significantly reducing the computational complexity. The algorithm yields an efficient solution to incremental LDA even when the number of classes as well as the set size is large. The incremental LDA method has been also shown useful for semi-supervised online learning. Label propagation is done by integrating the incremental LDA into an EM framework. The method has been demonstrated in the task of merging large datasets which were collected during MPEG standardization for face image retrieval, face authentication using the BANCA dataset, and object categorisation using the Caltech101 dataset.     

Inverse random under sampling for class imbalance problem and its application to multi-label classification

In this paper, a novel inverse random under sampling (IRUS) method is proposed for the class imbalance problem. The main idea is to severely under sample the majority class thus creating a large number of distinct training sets. For each training set we then find a decision boundary which separates the minority class from the majority class. By combining the multiple designs through fusion, we construct a composite boundary between the majority class and the minority class. The proposed methodology is applied on 22 UCI data sets and experimental results indicate a significant increase in performance when compared with many existing class-imbalance learning methods. We also present promising results for multi-label classification, a challenging research problem in many modern applications such as music, text and image categorization.     

Statistical properties of error estimators in performance assessment of recognition systems

The problem of estimating the error probability of a given classification system is considered. Statistical properties of the empirical error count (C) and the average conditional error (R) estimators are studied. It is shown that in the large sample case the R estimator is unbiased and its variance is less than that of the C estimator. In contrast to conventional methods of Bayes error estimation the unbiasedness of the R estimator for a given classifier can be obtained only at the price of an additional set of classified samples. On small test sets the R estimator may be subject to a pessimistic bias caused by the averaging phenomenon characterizing the functioning of conditional error estimators.     

Combining classifiers: A theoretical framework

The problem of classifier combination is considered in the context of the two main fusion scenarios: fusion of opinions based on identical and on distinct representations. We develop a theoretical framework for classifier combination for these two scenarios. For multiple experts using distinct representations we argue that many existing schemes such as the product rule, sum rule, min rule, max rule, majority voting, and weighted combination, can be considered as special cases of compound classification. We then consider the effect of classifier combination in the case of multiple experts using a shared representation where the aim of fusion is to obtain a better estimate of the appropriatea posteriori class probabilities. We also show that the two theoretical frameworks can be used for devising fusion strategies when the individual experts use features some of which are shared and the remaining ones distinct. We show that in both cases (distinct and shared representations), the expert fusion involves the computation of a linear or nonlinear function of thea posteriori class probabilities estimated by the individual experts. Classifier combination can therefore be viewed as a multistage classification process whereby thea posteriori class probabilities generated by the individual classifiers are considered as features for a second stage classification scheme. Most importantly, when the linear or nonlinear combination functions are obtained by training, the distinctions between the two scenarios fade away, and one can view classifier fusion in a unified way.