Connectivity enforced bayesian superpixels

Eker, Onur
In this study a general flow for clustering-based superpixel (SP) extraction methods is presented, while each step is analyzed in detail, and improvements are proposed. Considering general SP extraction method steps, initial grid alternatives are examined. The necessity of initial grid refinement is studied and unlike current approaches, a novel Edge Based Refinement step which does not break regular grid structure is proposed. Label update constraints are also analyzed in terms of preserving regular initial tiling, and Just Connected method enforcing connectivity from the beginning is proposed. The requirement of adjusting one of hyper-parameters, iteration count, for different image resolutions and different number of SPs is eliminated by determining the number of iterations relative to SP area. Considering these proposals,extensions to the state-of-the-art SP methods, SLIC+ and LASP+, are proposed which normalize spatial term with SP spatial covariance. Novel cost-functions SLIC++ and LASP++ are also presented for a further improvement with the normalization of spectral term with SP specific dynamic parameter. Finally, a Bayesian classifier is proposed for pixels during SP label assignment. Based on improvements in various steps mentioned above, a family of superpixel extraction methods including, SLIC++/R, SLIC++/H, LASP++/R. LASP++/H, BSP/R and BSP/H are presented. For the evaluation, a novel Boundary Achievable Segmentation Accuracy metric is proposed that replaces three frequent metrics from the literature. Compactness and area under curve approaches are also proposed as evaluation methods to minimize any performance ambiguity for the literature benchmarks. Both proposed spectral term and spatial term improvements significantly increase accuracy of generated SPs with no execution-time burden. In addition, employing Bayesian classifier leads to generate more accurate SPs in a shorter amount of run-time. Besides, with the proposed label update criteria, connectedness of SPs are ensured during generation process that preserves regular grid topology enabling them to be fed into conventional neural-networks.


Modeling and implementation of local volatility surfaces in Bayesian framework
Animoku, Abdulwahab; Uğur, Ömür; Yolcu-Okur, Yeliz (2018-06-01)
In this study, we focus on the reconstruction of volatility surfaces via a Bayesian framework. Apart from classical methods, such as, parametric and non-parametric models, we study the Bayesian analysis of the (stochastically) parametrized volatility structure in Dupire local volatility model. We systematically develop and implement novel mathematical tools for handling the classical methods of constructing local volatility surfaces. The most critical limitation of the classical methods is obtaining negativ...
Overconfidence and bubbles in experimental asset markets
Şahin, Serkan; Küçükkaya, Halit Engin; Yılmaz, Özlem; Department of Business Administration (2013)
The aim of this study is to investigate uncertainty levels of industries and explore those financial ratios that have the highest information content in determining the set of industry characteristics and use the most informative ratios selected in developing industry specific financial distress models. First, we employ factor analysis to determine the set of ratios that are most informative in specified industries. Second, we use entropy method as a Multiple Attribute Decision Making Model, to measure the ...
Dynamic modelling of a backhoe-loader
Kılıç, Boran; Balkan, Raif Tuna; Department of Mechanical Engineering (2009)
The aim of this study is to develop a dynamic model of the loader system of a backhoe-loader. Rigid bodies and joints in the loader mechanism and loader hydraulic system components are modelled and analyzed in the same environment using the physical modelling toolboxes inside the commercially available simulation software, MATLAB/Simulink. Interaction between the bodies and response of the hydraulic system are obtained by co-operating the mechanical and hydraulic analyses. System variables such as pressure,...
Meshless local Petrov-Galerkin method for plane elasticity problems
Erdayı, Deniz Can; Oral, Süha; Department of Mechanical Engineering (2014)
In this research, Meshless Local Petrov-Galerkin Method (MLPG) has been used in order to solve problems of elasto-statics. Moving least squares approximation (MLS) has been used to construct trial function. MLS weight function has been selected as test function. Thus, method is so-called MLPG1. Cantilever beam problem has been solved with MLPG. Effect of integration and influence domain sizes have been investigated for infinite plate with circular hole problem. Optimal parameters have been determined. Resul...
Monte Carlo model to assess effects of land-disposed E&P waste on groundwater
Parker, J.c.; Ünlü, Kahraman; Kemblowski, M.w. (1993-12-01)
The objective of this study was to develop quantitative stochastic methods to assess effects of land-disposed E&P waste on groundwater quality. A computer model was developed to assess the probability of groundwater contamination from salts and oily waste constituents disposed by landspreading or in waste pits. The model considers leaching of chemicals from the waste zone over time until the constituent mass is depleted. It considers volatilization of organics for landspreading. Water flow is controlled by ...
Citation Formats
O. Eker, “Connectivity enforced bayesian superpixels,” Thesis (M.S.) -- Graduate School of Natural and Applied Sciences. Electrical and Electronics Engineering., Middle East Technical University, 2019.