Browsing by Author "Turan, S."

Now showing 1 - 3 of 3

Circulating Insulin-Like Growth Factor Binding Protein-4 (igfbp-4) Is Not Regulated by Parathyroid Hormone and Vitamin D in Vivo: Evidence From Children With Rickets
(2010) Bereket, A.; Cesur, Y.; Özkan, B.; Adal, E.; Turan, S.; Onan, S.H.; Haklar, G.
Objective: Insulin-like growth factor binding protein-4 (IGFBP-4), inhibits IGF actions under a variety of experimental conditions. Parathyroid hormone (PTH), 1.25-hydroxy(OH)vitamin D, IGF-I, IGF-II and transforming growth factor (TGF)-β are the major regulators of IGFBP-4 production in vitro. However, little is known about the in vivo regulation of circulating IGFBP-4 in humans. Methods: We measured serum concentrations of calcium (Ca), phosphorus (P), alkaline phosphatase (ALP), PTH, vitamin D, IGF-I, IGFBP-3, and IGFBP-4 in infants (n=22) with nutritional rickets before and after treatment of rickets with vitamin D (300 000 U single dose po). Results: The mean±SD age of the patients was 1.3±1.6 years (range 0.2-3). Serum Ca and P increased, whereas ALP and PTH decreased after treatment (Ca from 6.6±1.4 to 9.5±1.6 mg/dL, P from 3.9±1.4 to 5.4±0.8 mg/dL, ALP from 2590±2630 to 1072±776 IU/mL and PTH from 407±248 to 27.4±20.8 ng/dL, respectively). Vitamin D levels were low (7.8±2.5 ng/mL) and increased after treatment (18.1±4.0 ng/mL, p<0.001). Serum IGF-I and IGFBP-3 levels both increased after treatment (IGF-I: 13.5±12.2 vs. 23.7±14.2 ng/mL, p<0.001 and IGFBP-3: 1108±544 vs. 1652±424 ng/mL, p<0.001). However, serum IGFBP-4 levels did not change significantly after treatment (18.8±8.0 vs. 21.5±4.8 ng/mL). No correlation between PTH and IGF-I, IGFBP-3 or IGFBP-4 was detected. Significant correlations were observed between PTH and ALP (r=0.53, p<0.05), and between IGF-I and IGFBP-3 (r=0.46, p<0.05). Conclusion: The results demonstrate that contrary to in vivo studies, circulating IGFBP-4 levels are not influenced by secondary hyperparathyroidism in vitamin D deficiency rickets since IGFBP-4 levels did not change after normalization of PTH with vitamin D treatment. © Journal of Clinical Research in Pediatric Endocrinology, Published by Galenos Publishing.
Emotion Analysis From Facial Expressions Using Convolutional Neural Networks
(Institute of Electrical and Electronics Engineers Inc., 2021) Coşkun Irmak, M.; Bilge Han Taş, M.; Turan, S.; Haşıloğlu, A.
In order to better understand human behavior, the emotional content of human facial expressions needs to be accurately analyzed and interpreted. While the perception of faces and facial expressions is a natural skill for humans, it still poses great challenges for computer systems. These difficulties result from the non-uniformity of the human face and differences in conditions such as lighting, shadows, face pose and orientation. Deep learning models, especially Convolutional Neural Networks (CNNs), have great potential to deal with these challenges due to their powerful automatic feature extraction and computational efficiency. In this study, a CNN model is proposed to classify seven different emotions (angry, disgust, fear, happy, sadness, surprise and neutral) using the FER-2013 dataset. With the proposed model, 70.62% accuracy on the training data and 70% on the test data has been achieved. The loss value was found to be 0.80 at the training stage and 0.86 at the testing stage. © 2021 IEEE
Real-Time Puddle Detection Using Convolutiona Neural Networks With Unmanned Aerial Vehicles
(Institute of Electrical and Electronics Engineers Inc., 2021) Bilge Han Taş, M.; Coşkun Irmak, M.; Turan, S.; Haşıloğlu, A.
The study was carried out in order to enable systems witli weak processing power and motion to detect objects using cloud services. In addition, the dataset is expanded by continuous labeling to create big data. In the study, it is aimed to detect objects using cloud-based deep learning methods with an unmanned aerial vehicle (UAV). In the study, training processes were carried out with Google Colaboratory, a cloud service provider. The training processes are a YOLO-based system, and a convolutional neural network was created by revising the parameters in line with the needs. The convolutional neural network model provides communication between neurons in the convolutional layers by bringing the image data to the desired pixel ranges. Unlabeled pictures are included in the training by being tagged. In this way, it is possible to continuously enlarge the data pool. Since the microcomputers used in UAVs are insufficient for these processes, a cloud-based training model has been created. As a result of the study, cloud-based deep learning models work as desired. It is possible to show the accuracy of the model with the low losses seen in the loss functions and the mAP value. Graphic cards with high processing power are needed to provide training. It is essential to use powerful graphics cards when working on image data. Cost reduced by using cloud services. The training was accelerated and high-rate object detections were made. YOLOv5x was used in the study. It is preferred because of its fast training and high frame rate. Recall 80% Precision 93% mAP 82.6% values were taken. © 2021 IEEE