Reinforcement learning (RL) is a promising tool for solving robust optimal well control problems ... more Reinforcement learning (RL) is a promising tool for solving robust optimal well control problems where the model parameters are highly uncertain and the system is partially observable in practice. However, the RL of robust control policies often relies on performing a large number of simulations. This could easily become computationally intractable for cases with computationally intensive simulations. To address this bottleneck, an adaptive multigrid RL framework is introduced which is inspired by principles of geometric multigrid methods used in iterative numerical algorithms. RL control policies are initially learned using computationally efficient low-fidelity simulations with coarse grid discretization of the underlying partial differential equations (PDEs). Subsequently, the simulation fidelity is increased in an adaptive manner towards the highest fidelity simulation that corresponds to the finest discretization of the model domain. The proposed framework is demonstrated using...
PURPOSE To evaluate the effect of the implantation of intracorneal ring segments (ICRS) in kerato... more PURPOSE To evaluate the effect of the implantation of intracorneal ring segments (ICRS) in keratoconus on the dynamic corneal response (DCR) parameters obtained with the Corvis ST (Oculus Optikgeräte GmbH). METHODS This prospective clinical study included patients who underwent ICRS implantation for keratoconus over a period of 1 year. On the day of the surgery and at least 1 month after ICRS implantation, the following measurements were made: corrected and uncorrected distance visual acuity, corneal tomography indices with the Pentacam (Oculus Optikgeräte GmbH), biomechanically corrected intraocular pressure (bIOP), Corvis ST DCR parameters, integrated inverse concave radius (1/R), deformation amplitude ratio (DA ratio), stiffness parameter at first applanation (SP-A1), stress-strain index (SSI), and highest concavity radius (HRC). RESULTS Fifty-one eyes of 40 patients were included with a median follow-up time of 3 months (interquartile ratio [IQR]: 2 to 6 months). Statistical analysis showed that ICRS implantation did not affect corneal biomechanical measurements evaluated with the Corvis ST, which was demonstrated by non-significant changes in the SP-A1 (P = .637), 1/R (P = .647), HRC (P = .177), DA ratio (P = .735), and SSI (P = .501). Additionally, the results showed that bIOP measurements were not significantly affected by ICRS implantation (P = .113). CONCLUSIONS ICRS implantation does not affect corneal biomechanical measurements in early follow-up. [J Refract Surg. 2022;38(4):264-269.].
The ideal smart home could be automatically controlled using a variety of electronic tools and de... more The ideal smart home could be automatically controlled using a variety of electronic tools and devices to perform everyday tasks. Smart home automation is crucially beneficial for human life, particularly when considering those with disabilities, inpatients, and elderly populations. In this paper, applications and systems for smart homes are investigated. During experimentation they were controlled via an Android mobile phone and the Arduino platform. Bluetooth Module HC-06 was used to connect the Arduino Uno R3 with the mobile phone. Five smart home applications were developed to control the lighting and electrical sockets, fan speed, temperature- and humidity-meter display/controls, as well as the fire-alarm and toxic-gas alarm systems. Herein, the definition, the graphical user interface, the required main components, and the control circuit connections are prepared and presented for each application. The graphical user interface was created using the RemoteXY website, which is a...
BackgroundCorneal refractive surgery has become reliable for correcting refractive errors, but it... more BackgroundCorneal refractive surgery has become reliable for correcting refractive errors, but it can induce unintended ocular changes that alter refractive outcomes. This study is to evaluate the unintended changes in ocular biometric parameters over a 6-month follow-up period after femtosecond laser-assisted laser in situ keratomileusis (FS-LASIK) and small incision lenticule extraction (SMILE).Methods156 consecutive myopic patients scheduled for FS-LASIK and SMILE were included in this study. Central corneal thickness (CCT), mean curvature of the corneal posterior surface (Kpm), internal anterior chamber depth (IACD) and the length from corneal endothelium to retina (ER) were evaluated before and after surgery over a 6-month period.ResultsBoth the FS-LASIK and SMILE groups (closely matched at the pre-surgery stage) experienced flatter Kpm, shallower IACD and decreased ER 1 week post-surgery (P < 0.01), and these changes were larger in FS-LASIK than in SMILE group. During the 1...
2019 Conference on Design and Architectures for Signal and Image Processing (DASIP), 2019
Convolutional Neural Networks (CNNs) have achieved significant success in image classification. O... more Convolutional Neural Networks (CNNs) have achieved significant success in image classification. One of the main reasons that CNNs achieve state-of-the-art accuracy is using many multi-scale learnable windowed feature detectors called kernels. Fetching of kernel feature weights from memory and performing the associated multiply and accumulate computations consume massive amount of energy. This hinders the widespread usage of CNNs, especially in embedded devices. In comparison with CNNs, decision forests are computationally efficient since they are composed of decision trees, which are binary classifiers by nature and can be implemented using AND-OR gates instead of costly multiply and accumulate units. In this paper, we investigate the migration of CNNs to decision forests as one of the promising approaches for reducing both execution time and power consumption while achieving acceptable accuracy. We introduce POLYCiNN, an architecture composed of a stack of decision forests. Each decision forest classifies one of the overlapped sub-images of the original image. Then, all decision forest classifications are fused together to classify the input image. In POLYCiNN, each decision tree is implemented in a single 6-input Look-Up Table and requires no memory access. Therefore, POLYCiNN can be efficiently mapped to simple and densely parallel hardware designs. We validate the performance of POLYCiNN on the benchmark image classification tasks of the MNIST, CIFAR-10 and SVHN datasets.
This study aims to estimate the reduction in collagen fibril density within the central 6 mm radi... more This study aims to estimate the reduction in collagen fibril density within the central 6 mm radius of keratoconic corneas through the processing of microstructure and videokeratography data. Collagen fibril distribution maps and topography maps were obtained for seven keratoconic and six healthy corneas, and topographic features were assessed to detect and calculate the area of the cone in each keratoconic eye. The reduction in collagen fibril density within the cone area was estimated with reference to the same region in the characteristic collagen fibril maps of healthy corneas. Together with minimum thickness and mean central corneal refractive power, the cone area was correlated with the reduction in the cone collagen fibrils. For the corneas considered, the mean area of keratoconic cones was 3.30 ± 1.90 mm 2 . Compared with healthy corneas, fibril density in the cones of keratoconic corneas was lower by as much as 35%, and the mean reduction was 17 ± 10%. A linear approximatio...
One of the main challenges in the parametrization of geological models is the ability to capture ... more One of the main challenges in the parametrization of geological models is the ability to capture complex geological structures often observed in the subsurface. In recent years, generative adversarial networks (GAN) were proposed as an efficient method for the generation and parametrization of complex data, showing state-of-the-art performances in challenging computer vision tasks such as reproducing natural images (handwritten digits, human faces, etc.). In this work, we study the application of Wasserstein GAN for the parametrization of geological models. The effectiveness of the method is assessed for uncertainty propagation tasks using several test cases involving different permeability patterns and subsurface flow problems. Results show that GANs are able to generate samples that preserve the multipoint statistical features of the geological models both visually and quantitatively. The generated samples reproduce both the geological structures and the flow statistics of the ref...
A. Elsheikh , D. Zhou, A. Eliasy and C. Boote School of Engineering, University of Liverpool, Liv... more A. Elsheikh , D. Zhou, A. Eliasy and C. Boote School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, elsheikh@liverpool.ac.uk NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, UK, elsheikh@liverpool.ac.uk School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, sgdzhou3@student.liverpool.ac.uk School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, eliasy.ashkan@gmail.com Structural Biophysics Group, School of Optometry and Vision Sciences, Cardiff University, BooteC@cardiff.ac.uk
A.Eliasy, K.Chen and A.Elsheikh School of Engineering, University of Liverpool, Liverpool L69 3GH... more A.Eliasy, K.Chen and A.Elsheikh School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, eliasy.ashkan@gmail.com School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, hskchen5@liverpool.ac.uk School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, elsheikh@liverpool.ac.uk NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, UK, elsheikh@liverpool.ac.uk
Reinforcement learning (RL) is a promising tool for solving robust optimal well control problems ... more Reinforcement learning (RL) is a promising tool for solving robust optimal well control problems where the model parameters are highly uncertain and the system is partially observable in practice. However, the RL of robust control policies often relies on performing a large number of simulations. This could easily become computationally intractable for cases with computationally intensive simulations. To address this bottleneck, an adaptive multigrid RL framework is introduced which is inspired by principles of geometric multigrid methods used in iterative numerical algorithms. RL control policies are initially learned using computationally efficient low-fidelity simulations with coarse grid discretization of the underlying partial differential equations (PDEs). Subsequently, the simulation fidelity is increased in an adaptive manner towards the highest fidelity simulation that corresponds to the finest discretization of the model domain. The proposed framework is demonstrated using...
PURPOSE To evaluate the effect of the implantation of intracorneal ring segments (ICRS) in kerato... more PURPOSE To evaluate the effect of the implantation of intracorneal ring segments (ICRS) in keratoconus on the dynamic corneal response (DCR) parameters obtained with the Corvis ST (Oculus Optikgeräte GmbH). METHODS This prospective clinical study included patients who underwent ICRS implantation for keratoconus over a period of 1 year. On the day of the surgery and at least 1 month after ICRS implantation, the following measurements were made: corrected and uncorrected distance visual acuity, corneal tomography indices with the Pentacam (Oculus Optikgeräte GmbH), biomechanically corrected intraocular pressure (bIOP), Corvis ST DCR parameters, integrated inverse concave radius (1/R), deformation amplitude ratio (DA ratio), stiffness parameter at first applanation (SP-A1), stress-strain index (SSI), and highest concavity radius (HRC). RESULTS Fifty-one eyes of 40 patients were included with a median follow-up time of 3 months (interquartile ratio [IQR]: 2 to 6 months). Statistical analysis showed that ICRS implantation did not affect corneal biomechanical measurements evaluated with the Corvis ST, which was demonstrated by non-significant changes in the SP-A1 (P = .637), 1/R (P = .647), HRC (P = .177), DA ratio (P = .735), and SSI (P = .501). Additionally, the results showed that bIOP measurements were not significantly affected by ICRS implantation (P = .113). CONCLUSIONS ICRS implantation does not affect corneal biomechanical measurements in early follow-up. [J Refract Surg. 2022;38(4):264-269.].
The ideal smart home could be automatically controlled using a variety of electronic tools and de... more The ideal smart home could be automatically controlled using a variety of electronic tools and devices to perform everyday tasks. Smart home automation is crucially beneficial for human life, particularly when considering those with disabilities, inpatients, and elderly populations. In this paper, applications and systems for smart homes are investigated. During experimentation they were controlled via an Android mobile phone and the Arduino platform. Bluetooth Module HC-06 was used to connect the Arduino Uno R3 with the mobile phone. Five smart home applications were developed to control the lighting and electrical sockets, fan speed, temperature- and humidity-meter display/controls, as well as the fire-alarm and toxic-gas alarm systems. Herein, the definition, the graphical user interface, the required main components, and the control circuit connections are prepared and presented for each application. The graphical user interface was created using the RemoteXY website, which is a...
BackgroundCorneal refractive surgery has become reliable for correcting refractive errors, but it... more BackgroundCorneal refractive surgery has become reliable for correcting refractive errors, but it can induce unintended ocular changes that alter refractive outcomes. This study is to evaluate the unintended changes in ocular biometric parameters over a 6-month follow-up period after femtosecond laser-assisted laser in situ keratomileusis (FS-LASIK) and small incision lenticule extraction (SMILE).Methods156 consecutive myopic patients scheduled for FS-LASIK and SMILE were included in this study. Central corneal thickness (CCT), mean curvature of the corneal posterior surface (Kpm), internal anterior chamber depth (IACD) and the length from corneal endothelium to retina (ER) were evaluated before and after surgery over a 6-month period.ResultsBoth the FS-LASIK and SMILE groups (closely matched at the pre-surgery stage) experienced flatter Kpm, shallower IACD and decreased ER 1 week post-surgery (P < 0.01), and these changes were larger in FS-LASIK than in SMILE group. During the 1...
2019 Conference on Design and Architectures for Signal and Image Processing (DASIP), 2019
Convolutional Neural Networks (CNNs) have achieved significant success in image classification. O... more Convolutional Neural Networks (CNNs) have achieved significant success in image classification. One of the main reasons that CNNs achieve state-of-the-art accuracy is using many multi-scale learnable windowed feature detectors called kernels. Fetching of kernel feature weights from memory and performing the associated multiply and accumulate computations consume massive amount of energy. This hinders the widespread usage of CNNs, especially in embedded devices. In comparison with CNNs, decision forests are computationally efficient since they are composed of decision trees, which are binary classifiers by nature and can be implemented using AND-OR gates instead of costly multiply and accumulate units. In this paper, we investigate the migration of CNNs to decision forests as one of the promising approaches for reducing both execution time and power consumption while achieving acceptable accuracy. We introduce POLYCiNN, an architecture composed of a stack of decision forests. Each decision forest classifies one of the overlapped sub-images of the original image. Then, all decision forest classifications are fused together to classify the input image. In POLYCiNN, each decision tree is implemented in a single 6-input Look-Up Table and requires no memory access. Therefore, POLYCiNN can be efficiently mapped to simple and densely parallel hardware designs. We validate the performance of POLYCiNN on the benchmark image classification tasks of the MNIST, CIFAR-10 and SVHN datasets.
This study aims to estimate the reduction in collagen fibril density within the central 6 mm radi... more This study aims to estimate the reduction in collagen fibril density within the central 6 mm radius of keratoconic corneas through the processing of microstructure and videokeratography data. Collagen fibril distribution maps and topography maps were obtained for seven keratoconic and six healthy corneas, and topographic features were assessed to detect and calculate the area of the cone in each keratoconic eye. The reduction in collagen fibril density within the cone area was estimated with reference to the same region in the characteristic collagen fibril maps of healthy corneas. Together with minimum thickness and mean central corneal refractive power, the cone area was correlated with the reduction in the cone collagen fibrils. For the corneas considered, the mean area of keratoconic cones was 3.30 ± 1.90 mm 2 . Compared with healthy corneas, fibril density in the cones of keratoconic corneas was lower by as much as 35%, and the mean reduction was 17 ± 10%. A linear approximatio...
One of the main challenges in the parametrization of geological models is the ability to capture ... more One of the main challenges in the parametrization of geological models is the ability to capture complex geological structures often observed in the subsurface. In recent years, generative adversarial networks (GAN) were proposed as an efficient method for the generation and parametrization of complex data, showing state-of-the-art performances in challenging computer vision tasks such as reproducing natural images (handwritten digits, human faces, etc.). In this work, we study the application of Wasserstein GAN for the parametrization of geological models. The effectiveness of the method is assessed for uncertainty propagation tasks using several test cases involving different permeability patterns and subsurface flow problems. Results show that GANs are able to generate samples that preserve the multipoint statistical features of the geological models both visually and quantitatively. The generated samples reproduce both the geological structures and the flow statistics of the ref...
A. Elsheikh , D. Zhou, A. Eliasy and C. Boote School of Engineering, University of Liverpool, Liv... more A. Elsheikh , D. Zhou, A. Eliasy and C. Boote School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, elsheikh@liverpool.ac.uk NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, UK, elsheikh@liverpool.ac.uk School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, sgdzhou3@student.liverpool.ac.uk School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, eliasy.ashkan@gmail.com Structural Biophysics Group, School of Optometry and Vision Sciences, Cardiff University, BooteC@cardiff.ac.uk
A.Eliasy, K.Chen and A.Elsheikh School of Engineering, University of Liverpool, Liverpool L69 3GH... more A.Eliasy, K.Chen and A.Elsheikh School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, eliasy.ashkan@gmail.com School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, hskchen5@liverpool.ac.uk School of Engineering, University of Liverpool, Liverpool L69 3GH, UK, elsheikh@liverpool.ac.uk NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, UK, elsheikh@liverpool.ac.uk
Uploads