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.
The research deals with the interchange of the sign transformed from the universal to the local in the theatrical show through the direction processing in the production of a communicative artistic discourse and message, thus making the process of reading the speech and recognizing it by taking into account the cultural differences, customs and local rituals of each country, region, or area. The problem of the research was focused on answering the following question: What are the requirements for the sign in terms of its transformation between the universality and locality in the read-out?
The importance of research is to determine the requiremen
... Show MoreConducted Althilelat chemical models of crude oil back to the reservoir Fertile from the fields of Baghdad and Kut and models of crude oil back to the reservoir ??????? of Haklbe Tikrit and Baghdad were calculated their properties Alvezaúah Kalkthaqh and weight, quality and degree of August j (API) and know the quality Nfothma that was light or heavy and make the comparison between Alinvtin also conducted chemical analyzes of the two models of Almia associated with each of the oil above Almkmnin and measured Ktvthma and Zojithma and concentrations of some dissolved salts in them and clarify the relationship between the oil reservoir and water associated with oil fields...
More than 450 distinct types of human papilloma virus recognized via recent molecular techniques. The low and high oncogenic risk-HPV genotypes have an association with a variety of benign and malignant tumors in the oropharyngeal and nasopharyngeal localizations. This study aimed to determine the rate of DNA detection of HPV genotype 6/11 in non-oncologic nasopharyngeal and palatine tonsillar tissues from pediatric patients subjected to adeno-tonsillectomies. A total number of 64 tissue specimens enrolled; 44 non-oncologic nasopharyngeal and palatine hypertrophied tissue specimens from 22 pediatric patients sustained combined adeno-tonsillectomies and compared to 20 nasal trimmed tissues with unremarkable pathological changes (in
... Show MoreBackground Doxorubicin (DOX) is a potent antineoplastic agent used in treating various adult and pediatric cancers, but it tends to provoke dose-dependent cardiotoxicity. Ezetimibe (EZE), a cholesterol-lowering drug, has been reported to possess defensive actions against oxidative stress and inflammation, which are two of the main proposed mechanisms underlying the development of DOX-induced cardiotoxicity (DIC), hence, we aimed to inspect the possible protective effect of EZE against DIC in rats. Methods 24 adult male Wistar rats were allocated into four groups of six: control, DOX, 10 mg/kg EZE plus DOX and 20 mg/kg EZE plus DOX. At the end of the study, the experimental rats were anesthetized and blood samples were collected for b
... Show MoreObjective: Determination the effectiveness of educational program on female students’ practices toward premenstrual.
Methodology: A quasi-experimental design study was conducted involving (140) student purposely in four secondary schools at Al-sadder city (70) student for study group and (70) for control group. The prevalence of PMS selected through American College of Obstetricians and Gynecologists (ACOG) (2015) criteria to select PMS students before program. The education program were set in four steps, the first step (pre-test) is to assess the practices, before the implementation of the program, the second step is implementing the program, following two steps post-test I and II betwe
... Show MoreAchieving an accurate and optimal rate of penetration (ROP) is critical for a cost-effective and safe drilling operation. While different techniques have been used to achieve this goal, each approach has limitations, prompting researchers to seek solutions. This study’s objective is to conduct the strategy of combining the Bourgoyne and Young (BYM) ROP equations with Bagging Tree regression in a southern Iraqi field. Although BYM equations are commonly used and widespread to estimate drilling rates, they need more specific drilling parameters to capture different ROP complexities. The Bagging Tree algorithm, a random forest variant, addresses these limitations by blending domain kno