In this thesis,we study how a meta-analysis of genetic association studies is influenced by the degree of genotype imputation uncertainty in the studies combined and the size of meta-analysis. We consider the fixed effect meta-analysis model to evaluate the accuracy and efficiency of imputation-based meta-analysis results under different levels of imputation accuracy. We also examine the impact of genotype imputation on the between-study heterogeneity and type 1 error in the random effects meta-analysis model. Simulation results reaffirm that meta-analysis boosts the power of detecting genetic associations compared to individual study results. However, the power deteriorates with increasing uncertainty in imputed genotypes. Genotype imputation affects a random effects meta-analysis in a non-obvious way as estimation of between-study heterogeneity and interpretation of association results depend heavily on the number of studies combined. We propose an adjusted fixed effect meta-analysis approach for adding imputation-based studies to a meta-analysis of existing typed studies in a controlled way to improve precision and reliability. The proposed method should help in designing an effective meta-analysis study.

In recent years nickel is mostly produced from lateritic ore deposits such as nontronite, limonite, etc. Resource estimation is difficult for laterite deposits as they have a weak and heterogeneous form. 3D modeling software are rather suitable for deposits having tabular or vein type ores. In this study the most appropriate estimation technique for resource estimation of nickel laterite deposits was investigated. One of the known nickel laterite deposits in Turkey is located at T&uuml / rkmen&ccedil / ardagi - G&ouml / rdes region. Since the nickel (Ni) grade recovered from drilling studies seem to be very low, a reconciliation pit having dimensions of 40 m x 40 m x 15 m in x-y-z directions was planned by Meta Nikel Kobalt Mining Company (META), the license owner of the mine, to produce nickel ore. 13 core drilling and 13 reverse circulation drilling (RC) and 26 column samplings adjacent to each drillholes were located in this area. Those three sampling results were compared to each other and as well as the actual production values obtained from reconciliation pit. On the other side 3D computer modeling was also used to model the nickel resource in T&uuml / rkmen&ccedil / ardagi - G&ouml / rdes laterites. The results obtained from both inverse distance weighting and kriging methods were compared to the results of actual production to find out the applicability of 3D modeling to laterite deposits. Modeling results showed that Ni grade of the reconciliation pit in T&uuml / rkmen&ccedil / ardagi - G&ouml / rdes, considering 0.5% Ni cut-off value, by using drillholes data, inverse distance weighting method estimates 622 tonnes with 0.553% Ni and kriging method estimates 749 tonnes with 0.527% Ni. The actual production pit results provided 4,882 tonnes of nickel ore with 0.649% Ni grade. These results show that grade values seem to be acceptable but in terms of tonnage, there are significant differences between theoretical estimated values and production values.

With the availability of inexpensive hardware and software, digital imaging has become an important medium of communication in our daily lives. A huge amount of digital images are being collected and become available through the internet and stored in various fields such as personal image collections, medical imaging, digital arts etc. Therefore, it is important to make sure that images are stored, searched and accessed in an efficient manner. The use of bag of visual words (BOW) model for modelling images based on local invariant features computed at interest point locations has become a standard choice for many computer vision tasks. Based on this promising model, this thesis investigates three main problems: natural scene classification, annotation and retrieval. Given an image, the task is to design a system that can determine to which class that image belongs to (classification), what semantic concepts it contain (annotation) and what images are most similar to (retrieval). This thesis contributes to scene classification by proposing a weighting approach, named keypoints density-based weighting method (KDW), to control the fusion of colour information and bag of visual words on spatial pyramid layout in a unified framework. Different configurations of BOW, integrated visual vocabularies and multiple image descriptors are investigated and analyzed. The proposed approaches are extensively evaluated over three well-known scene classification datasets with 6, 8 and 15 scene categories using 10-fold cross validation. The second contribution in this thesis, the scene annotation task, is to explore whether the integrated visual vocabularies generated for scene classification can be used to model the local semantic information of natural scenes. In this direction, image annotation is considered as a classification problem where images are partitioned into 10x10 fixed grid and each block, represented by BOW and different image descriptors, is classified into one of predefined semantic classes. An image is then represented by counting the percentage of every semantic concept detected in the image. Experimental results on 6 scene categories demonstrate the effectiveness of the proposed approach. Finally, this thesis further explores, with an extensive experimental work, the use of different configurations of the BOW for natural scene retrieval.

With the availability of inexpensive hardware and software, digital imaging has become an important medium of communication in our daily lives. A huge amount of digital images are being collected and become available through the internet and stored in various fields such as personal image collections, medical imaging, digital arts etc. Therefore, it is important to make sure that images are stored, searched and accessed in an efficient manner. The use of bag of visual words (BOW) model for modelling images based on local invariant features computed at interest point locations has become a standard choice for many computer vision tasks. Based on this promising model, this thesis investigates three main problems: natural scene classification, annotation and retrieval. Given an image, the task is to design a system that can determine to which class that image belongs to (classification), what semantic concepts it contain (annotation) and what images are most similar to (retrieval). This thesis contributes to scene classification by proposing a weighting approach, named keypoints density-based weighting method (KDW), to control the fusion of colour information and bag of visual words on spatial pyramid layout in a unified framework. Different configurations of BOW, integrated visual vocabularies and multiple image descriptors are investigated and analyzed. The proposed approaches are extensively evaluated over three well-known scene classification datasets with 6, 8 and 15 scene categories using 10-fold cross validation. The second contribution in this thesis, the scene annotation task, is to explore whether the integrated visual vocabularies generated for scene classification can be used to model the local semantic information of natural scenes. In this direction, image annotation is considered as a classification problem where images are partitioned into 10x10 fixed grid and each block, represented by BOW and different image descriptors, is classified into one of predefined semantic classes. An image is then represented by counting the percentage of every semantic concept detected in the image. Experimental results on 6 scene categories demonstrate the effectiveness of the proposed approach. Finally, this thesis further explores, with an extensive experimental work, the use of different configurations of the BOW for natural scene retrieval. / Applied Science University in Jordan