Human pose estimation is considered one of the major challenges in the field of Computer Vision, playing an integral role in a large variety of technology domains. While, in the last few years, there ...has been an increased number of research approaches towards CNN-based 2D human pose estimation from RGB images, respective work on CNN-based 3D human pose estimation from depth/3D data has been rather limited, with current approaches failing to outperform earlier methods, partially due to the utilization of depth maps as simple 2D single-channel images, instead of an actual 3D world representation. In order to overcome this limitation, and taking into consideration recent advances in 3D detection tasks of similar nature, we propose a novel fully-convolutional, detection-based 3D-CNN architecture for 3D human pose estimation from 3D data. The architecture follows the sequential network architecture paradigm, generating per-voxel likelihood maps for each human joint, from a 3D voxel-grid input, and is extended, through a bottom-up approach, towards multi-person 3D pose estimation, allowing the algorithm to simultaneously estimate multiple human poses, without its runtime complexity being affected by the number of people within the scene. The proposed multi-person architecture, which is the first within the scope of 3D human pose estimation, is comparatively evaluated on three single person public datasets, achieving state-of-the-art performance, as well as on a public multi-person dataset achieving high recognition accuracy.
•3D-CNN architecture for 3D human pose estimation from 3D/depth data.•Fully-convolutional, detection-based, sequential network architecture.•Extension, through a bottom-up approach, towards multi-person 3D pose estimation.•Runtime complexity not affected by number of people in the scene.•SoA performance in 3 single person datasets and a novel multi-person dataset.
Kerogen is a microporous amorphous solid, which is the major component of the organic matter scattered in the potentially lucrative shale formations hosting shale gas. A deeper understanding of the ...way kerogen porosity characteristics affect the transport properties of hosted gas is important for the optimal design of the extraction process. In this work, we employ molecular simulation techniques to investigate the role of porosity on the adsorption and transport behavior of shale gas in overmature type II kerogen found in many currently productive shales. To account for the wide range of porosity characteristics present in the real system, a large set of 60 kerogen structures that exhibit a diverse set of void space attributes was used. Grand canonical Monte Carlo simulations were performed for the study of the adsorption of CH4, C2H6, n-C4H10, and CO2 at 298.15 and 398.15 K and a variety of pressures. The amount adsorbed is found to correlate linearly with the porosity of the kerogen. Furthermore, the adsorption of a quaternary mixture of CH4, C2H6, CO2, and N2 was investigated under the same conditions, indicating that a composition resembling that of the shale gas is achieved under higher temperature and pressure values, i.e., conditions closer to those prevailing in the hosting shale field. The diffusion of CH4, C2H6, and CO2, both as pure components and as components of the quaternary mixture, was investigated using equilibrium molecular dynamics simulations at temperatures of 298.15 and 398.15 K and pressures of 1 and 250 atm. In addition to the effect of temperature and pressure, the importance of limiting pore diameter (LPD), maximum pore diameter (MPD), accessible volume (V acc), and accessible surface (S acc) on the observed adsorbed amount and diffusion coefficient was revealed by qualitative relationships. The diffusion across the models was found to be anisotropic and the maximum component of the diffusion coefficient to correlate linearly with LPD, indicating that the controlling step of the transport process is the crossing of the limiting pore region. Finally, the transport behavior of the pure compounds was compared with their transport properties when in mixture and it was found that the diffusion coefficient of each compound in the mixture is similar to the corresponding one under pure conditions. This observation agrees with earlier studies in different kerogen models comprising wider pores that have revealed negligible cross-correlation Onsager coefficients.
Organic molecules can crystallize in multiple structures or polymorphs, yielding crystals with very different physical and mechanical properties. The prediction of the polymorphs that may appear in ...nature is a challenge with great potential benefits for the development of new products and processes. A multistage crystal structure prediction (CSP) methodology is applied to axitinib, a pharmaceutical molecule with significant polymorphism arising from molecular flexibility. The CSP study is focused on those polymorphs with one molecule in the asymmetric unit. The approach successfully identifies all four known polymorphs within this class, as well as a large number of other low-energy structures. The important role of conformational flexibility is highlighted. The performance of the approach is discussed in terms of both the quality of the results and various algorithmic and computational aspects, and some key priorities for further work in this area are identified.
Display omitted
•First ab initio prediction of the polymorphs of axitinib, a challenging pharmaceutical molecule.•Successful application of a systematic procedure to identify polymorphs with one molecule in asymmetric unit.•Analysis of strengths and weaknesses of state of the art for crystal structure prediction.•Identification of future research needs and priorities in this area.
Natural gas production from shale formations is one of the most recent and fast growing developments in the oil and gas industry. The accurate prediction of the adsorption and transport of shale gas ...is essential for estimating shale gas production capacity and improving existing extractions. To realistically represent heterogeneous shale formations, a composite pore model was built from a kaolinite slit mesopore hosting a kerogen matrix. Moreover, empty slabs (2, 3, and 4 nm) were added between the kerogen matrix and siloxane surface of kaolinite. Using Grand–Canonical Monte Carlo (GCMC) and molecular dynamics (MD) simulations, the adsorption and diffusion of pure methane, pure ethane, and a shale gas mixture were computed at various high pressures (100, 150, and 250 atm) and temperature of 298.15 K. The addition of an inner slit pore was found to significantly increase the excess adsorption of methane, as a pure component and in the shale gas mixture. The saturation of the composite pore with methane was observed to be at a higher pressure compared to ethane. The excess adsorption of carbon dioxide was not largely affected by pressure, and the local number density profile showed its strong affinity to kerogen micropores and the hydroxylated gibbsite surface under all conditions and pore widths. Lateral diffusion coefficients were found to increase with increasing the width of the empty slab inside the composite pore. Statistical errors of diffusion coefficients were found to be large for the case of shale gas components present at low composition. A larger composite pore configuration was created to investigate the diffusion of methane in different regions of the composite pore. The calculated diffusion coefficients and mean residence times were found to be indicative of the different adsorption mechanisms occurring inside the pore.
The adsorption behavior inside kaolinite mesopores of aqueous solutions of various salts and additives is investigated using Molecular Dynamics simulations. In particular, we examine the various ...combinations of water + salt, water + additive, and water + salt + additive mixtures, where the salts are NaCl, CsCl, SrCl2, and RaCl2 and the additives are methanol and citric acid. Citric acid is modeled in two forms, namely, fully protonated (H3A) and fully deprotonated (A3–), the latter being prevalent in neutral pH conditions, in accordance with the kaolinite structure employed. The force fields used for the individual system components include CLAYFF for the kaolinite mesopores, SPC/E for water, parameters optimized for the SPC/E water model based on hydration free energies (HFE) for ions, and general Amber force field (GAFF) for the additives. The spatial distributions along the kaolinite pore are delineated and reveal the preferential adsorption behavior of the various species with respect to the gibbsite and siloxane surface, as well as the effect on this behavior of the interactions between the various species. Furthermore, we examine the hydrogen bonds formed between the kaolinite surfaces and water molecules as well as the additives. For the case of citric acid, which tends to aggregate, a cluster analysis is also carried out, in order to examine the effect of the various ions on the cluster formation. Finally, through the calculation of lateral diffusion coefficients and mean residence times, we provide insights on the mobility of the various species inside the kaolinite mesopores.
Shale gas is an unconventional source of energy, which has attracted a lot of attention during the last years. Kerogen is a prime constituent of shale formations and plays a crucial role in shale gas ...technology. Significant experimental effort in the study of shales and kerogen has produced a broad diversity of experimentally determined structural and thermodynamic properties even for samples of the same well. Moreover, proposed methods reported in the literature for constructing realistic bulk kerogen configurations have not been thoroughly investigated. One of the most important characteristics of kerogens is their porosity, due to its direct connection with their transport properties and its potential as discriminating and classifying metric between samples. In this study, molecular dynamics (MD) simulations are used to study the porosity of model kerogens. The porosity is controlled effectively with systematic variations of the number and the size of dummy LJ particles that are used during the construction of system’s configuration. The porosity of each sample is characterized with a newly proposed algorithm for analyzing the free space of amorphous materials. It is found that, with moderately sized configurations, it is possible to construct percolated pores of interest in the shale gas industry.
This thesis deals with the investigation of novel techniques for human pose estimation (HPE) using sparse depth/3D data, in order to develop a standalone, high-accuracy, low-latency human pose ...estimation module, suitable for deployment in systems with limited processing resources. Based on the existing work and motivated by the significant progress that has been achieved in the relevant fields, two novel methods for the estimation and tracking of the human pose utilising sparse depth/3D data, are proposed. First, a real-time human pose estimation and tracking framework is developed, which builds upon an already established human-template-tracking based approach, utilising the 3D Signed Distance Function (SDF) data representation. A series of complementary tracking features are introduced, tackling specifically the issues of free space violation, body part visibility and leg intersection, which are typically encountered under real-life monitoring conditions. The method is experimentally evaluated on a series of publicly available datasets, achieving state-of-the-art (SOA) performance, while also successfully utilised for human behavioural modelling on an autonomous robotic platform. Due to inherent limitations of this tracking-based approach, such as the requirement for clearly segmented human/background data and the use of an out-of-the-box initialiser, a second, deep learning-based architecture is investigated. Specifically, a detection-based 3D-CNN architecture for 3D human pose estimation from 3D data is introduced, following the sequential network architecture paradigm. It utilises a volumetric data representation, and generates 3D heatmaps corresponding to potential locations of the human joints in the scene, achieving state-of-the-art accuracy. Additionally, a 3D body-part detector is incorporated, extending the architecture towards multi-person 3D pose estimation, the first such method for 3D data. However, the 3D CNN architecture comes at a steep computational cost, making it unsuitable for implementation on low power systems. Thus, the final contribution of this thesis includes the investigation of computationally efficient 3D CNN design guidelines, in order to reduce the computational complexity of the developed model. The result of this investigation is a novel 3D-CNN architecture for multi-person pose estimation from 3D data, composed mainly of 3D depthwise residual bottleneck units, SE blocks and a decomposed strided input layer. This optimised version performs comparably to SOA methods on two public datasets, while requiring significantly fewer computational resources and achieving a speedup of over 100x on a modern low power mobile device, and a reduction in model size of approximately 50x.
The sixth blind test of organic crystal structure prediction (CSP) methods has been held, with five target systems: a small nearly rigid molecule, a polymorphic former drug candidate, a chloride salt ...hydrate, a co‐crystal and a bulky flexible molecule. This blind test has seen substantial growth in the number of participants, with the broad range of prediction methods giving a unique insight into the state of the art in the field. Significant progress has been seen in treating flexible molecules, usage of hierarchical approaches to ranking structures, the application of density‐functional approximations, and the establishment of new workflows and `best practices' for performing CSP calculations. All of the targets, apart from a single potentially disordered Z′ = 2 polymorph of the drug candidate, were predicted by at least one submission. Despite many remaining challenges, it is clear that CSP methods are becoming more applicable to a wider range of real systems, including salts, hydrates and larger flexible molecules. The results also highlight the potential for CSP calculations to complement and augment experimental studies of organic solid forms.
The results of the sixth blind test of organic crystal structure prediction methods are presented and discussed, highlighting progress for salts, hydrates and bulky flexible molecules, as well as on‐going challenges.
Personal assistive robots to be realized in the near future should have the ability to seamlessly coexist with humans in unconstrained environments, with the robot’s capability to understand and ...interpret the human behavior during human–robot cohabitation significantly contributing towards this end. Still, the understanding of human behavior through a robot is a challenging task as it necessitates a comprehensive representation of the high-level structure of the human’s behavior from the robot’s low-level sensory input. The paper at hand tackles this problem by demonstrating a robotic agent capable of apprehending human daily activities through a method, the Interaction Unit analysis, that enables activities’ decomposition into a sequence of units, each one associated with a behavioral factor. The modelling of human behavior is addressed with a Dynamic Bayesian Network that operates on top of the Interaction Unit, offering quantification of the behavioral factors and the formulation of the human’s behavioral model. In addition, light-weight human action and object manipulation monitoring strategies have been developed, based on RGB-D and laser sensors, tailored for onboard robot operation. As a proof of concept, we used our robot to evaluate the ability of the method to differentiate among the examined human activities, as well as to assess the capability of behavior modeling of people with Mild Cognitive Impairment. Moreover, we deployed our robot in 12 real house environments with real users, showcasing the behavior understanding ability of our method in unconstrained realistic environments. The evaluation process revealed promising performance and demonstrated that human behavior can be automatically modeled through Interaction Unit analysis, directly from robotic agents.
We investigate the ability of current ab initio crystal structure prediction techniques to identify the polymorphs of 5‐methyl‐2‐(2‐nitrophenyl)amino‐3‐thiophenecarbonitrile, also known as ROY ...because of the red, orange and yellow colours of its polymorphs. We use a methodology combining the generation of a large number of structures based on a computationally inexpensive model using the CrystalPredictor global search algorithm, and the further minimization of the most promising of these structures using the CrystalOptimizer local minimization algorithm which employs an accurate, yet efficiently constructed, model based on isolated‐molecule quantum‐mechanical calculations. We demonstrate that this approach successfully predicts the seven experimentally resolved structures of ROY as lattice‐energy minima, with five of these structures being within the 12 lowest energy structures predicted. Some of the other low‐energy structures identified are likely candidates for the still unresolved polymorphs of this molecule. The relative stability of the predicted structures only partially matches that of the experimentally resolved polymorphs. The worst case is that of polymorph ON, whose relative energy with respect to Y is overestimated by 6.65 kJ mol−1. This highlights the need for further developments in the accuracy of the energy calculations.