Clinical keratoconus (KCN) detection is a challenging and time-consuming task. In the diagnosis process, ophthalmologists must revise demographic and clinical ophthalmic examinations. The latter include slit-lamb, corneal topographic maps, and Pentacam indices (PI). We propose an Ensemble of Deep Transfer Learning (EDTL) based on corneal topographic maps. We consider four pretrained networks, SqueezeNet (SqN), AlexNet (AN), ShuffleNet (SfN), and MobileNet-v2 (MN), and fine-tune them on a dataset of KCN and normal cases, each including four topographic maps. We also consider a PI classifier. Then, our EDTL method combines the output probabilities of each of the five classifiers to obtain a decision based on the fusion of probabilities. Individually, the classifier based on PI achieved 93.1% accuracy, whereas the deep classifiers reached classification accuracies over 90% only in isolated cases. Overall, the average accuracy of the deep networks over the four corneal maps ranged from 86% (SfN) to 89.9% (AN). The classifier ensemble increased the accuracy of the deep classifiers based on corneal maps to values ranging (92.2% to 93.1%) for SqN and (93.1% to 94.8%) for AN. Including in the ensemble-specific combinations of corneal maps’ classifiers and PI increased the accuracy to 98.3%. Moreover, visualization of first learner filters in the networks and Grad-CAMs confirmed that the networks had learned relevant clinical features. This study shows the potential of creating ensembles of deep classifiers fine-tuned with a transfer learning strategy as it resulted in an improved accuracy while showing learnable filters and Grad-CAMs that agree with clinical knowledge. This is a step further towards the potential clinical deployment of an improved computer-assisted diagnosis system for KCN detection to help ophthalmologists to confirm the clinical decision and to perform fast and accurate KCN treatment.
Reverse Osmosis (RO) has already proved its worth as an efficient treatment method in chemical and environmental engineering applications. Various successful RO attempts for the rejection of organic and highly toxic pollutants from wastewater can be found in the literature over the last decade. Dimethylphenol is classified as a high-toxic organic compound found ubiquitously in wastewater. It poses a real threat to humans and the environment even at low concentration. In this paper, a model based framework was developed for the simulation and optimisation of RO process for the removal of dimethylphenol from wastewater. We incorporated our earlier developed and validated process model into the Species Conserving Genetic Algorithm (SCG
... Show MoreThis study used deep eutectic solvent (DES) as the liquid membrane in a bulk liquid membrane system (BLM) to remove glycerol from waste cooking oil‐based biodiesel. The DES was prepared from choline chloride and tetraethylene glycol at a molar ratio of 1:5. Diethyl ether was employed as a novel strip phase for the glycerol in BLM. The effects of the DES: biodiesel ratio, stirring speed, and extraction time on the extraction and stripping efficiencies were investigated. The results showed that BLM could give better glycerol removal from biodiesel than mechanical shaking. Increasing the DES: biodiesel ratio, stirring speed, and extraction time can enhance glycerol removal from the feed phase, achievi
This research aims to investigate the thermal performance of different thermal composite insulators, wrapped around a closed-loop copper pipe (CLP). To achieve this aim a system was designed and manufactured. It is consisted of closed water tank insulated by Rock Wool, and supplied with two electric heaters, two thermostat, a flow meter, a water pump, digital temperature scales, and four series of (CLP).
Six insulators were prepared namely; composites of Impregnated Fiberglass with Elastoclad and foaming Rubber (FER), Impregnated Fiberglass with Elastoclad resin and Polymeric Membrane (FEM), Impregnated Fiberglass with Polyurethane thermoset resin and Foaming Rubber (FUR), Impregnated Fiberglass with Polyurethane thermoset resin and P
Ionic liquids (ILs) and deep eutectic solvents (DESs) have been found to be highly effective as electrolytes in TiO2 NTAs-graphite cells when combined with additives that enhance conductivity by reducing the viscosity of these liquids. The presence of CaCl2.6H2O: Acetamide DES with DI water as an additive resulted in a cell voltage of 1.31V and an internal resistance of 19 ohm. This can be attributed to the concentration and quality of the ionic species. The cells exhibited an interesting response to the AlCl3-chloroacetamide IL with dichloromethane DCM as an additive, with a cell voltage of 1.81V and an internal resistance of 5.0 ohm. Once again, this is influenced by the quality and concentration of the ionic species. Furthermore,
... Show MoreThe transfer function model the basic concepts in the time series. This model is used in the case of multivariate time series. As for the design of this model, it depends on the available data in the time series and other information in the series so when the representation of the transfer function model depends on the representation of the data In this research, the transfer function has been estimated using the style nonparametric represented in two method local linear regression and cubic smoothing spline method The method of semi-parametric represented use semiparametric single index model, With four proposals, , That the goal of this research is comparing the capabilities of the above mentioned m
... Show MoreThe lethality of inorganic arsenic (As) and the threat it poses have made the development of efficient As detection systems a vital necessity. This research work demonstrates a sensing layer made of hydrous ferric oxide (Fe2H2O4) to detect As(III) and As(V) ions in a surface plasmon resonance system. The sensor conceptualizes on the strength of Fe2H2O4 to absorb As ions and the interaction of plasmon resonance towards the changes occurring on the sensing layer. Detection sensitivity values for As(III) and As(V) were 1.083 °·ppb−1 and 0.922 °·ppb