Parametric family of statistical distributions are of great importance for several applications.In particular, we propose to investigate the generalized Gamma mixture model (g MM) for modeling and classifying medical imaging (Chest x-ray and CT-scans). The main advantage of this mixture over some existing Gaussian models is that it allows additional flexibility in shape modeling, which is crucial for classification systems. In order to capture accurately the intrinsic nature of medical images, we propose to derive some efficient measures based on Fisher, Kullback-Leibler and Bhattacharyya distances for the mixtures of generalized Gamma distributions. Indeed, the main idea is to investigate these distances effectively via the statistical model parameters in order to make our proposed scheme particularly appropriate for image classification problem. The proposed approach involves the extraction of robust texture descriptors, the learning of mixture model g MM via the expectation-maximization (EM) and Newton-Raphson algorithms, and the classification of images using the derived mixtures-based distances. We evaluate our model against the challenging problem of early diagnosis of pneumonia diseases. Experimental results on different datasets show the merits of our developed framework compared with the other methods.

No abstract

In this paper, a novel statistical generative model based on hierarchical Pitman-Yor process and generalized Dirichlet distributions (GDs) is presented. The proposed model allows us to perform joint clustering and feature selection thanks to the interesting properties of the GD distribution. We develop an online variational inference algorithm, formulated in terms of the minimization of a Kullback-Leibler divergence, of our resulting model that tackles the problem of learning from high-dimensional examples. This variational Bayes formulation allows simultaneously estimating the parameters, determining the model's complexity, and selecting the appropriate relevant features for the clustering structure. Moreover, the proposed online learning algorithm allows data instances to be processed in a sequential manner, which is critical for large-scale and real-time applications. Experiments conducted using challenging applications, namely, scene recognition and video segmentation, where our approach is viewed as an unsupervised technique for visual learning in high-dimensional spaces, showed that the proposed approach is suitable and promising.

No abstract