D-optimal chemical balance weighing designs with diagonal covariance matrix of experimental errors

SummaryIn this paper, some problems related to determining experimental plans satisfying the criterion of D-optimality are presented. Moreover, the optimality conditions and relations between the parameters of the chemical balance weighing designs are described, and some construction examples are given.

Object recognition using hybrid boosting method

Li (ICCV, 2005) proposed a novel generative/discriminative way to combine features with different types and use them to learn labels in the images. However, the mixture of Gaussian used in Li’s paper suffers greatly from the curse of dimensionality. Here I propose an alternative approach to generate local region descriptor. I treat GMM with diagonal covariance matrix and PCA as separate features, and combine them as the local descriptor. In this way, we could reduce the computational time for mixture model greatly while score greater 90% accuracies for caltech-4 image sets.

Object recognition using hybrid boosting method

Li (ICCV, 2005) proposed a novel generative/discriminative way to combine features with different types and use them to learn labels in the images. However, the mixture of Gaussian used in Li’s paper suffers greatly from the curse of dimensionality. Here I propose an alternative approach to generate local region descriptor. I treat GMM with diagonal covariance matrix and PCA as separate features, and combine them as the local descriptor. In this way, we could reduce the computational time for mixture model greatly while score greater 90% accuracies for caltech-4 image sets.

Relations between regular A-optimal chemical and spring balance weighing designs with diagonal covariance matrix of errors

Summary In this paper, we study the relationships between regular A-optimal spring balance weighing designs and regular A-optimal chemical balance weighing designs. We give the basic relation between these designs in the case where the errors are uncorrelated and they have different variances. We give some examples of methods of construction of such designs.