• Journal of Periodontal and Implant Science
• /
• v.51 no.5
• /
• pp.342-351
• /
• 2021

Purpose: The aim of this study was to compare the inflammatory and lipid profile of patients with and without peri-implantitis. Methods: A cross-sectional biochemical study was carried out in which blood samples were collected from 16 patients with peri-implantitis and from 31 subjects with healthy implants. Clinical peri-implant parameters were obtained from all subjects. Levels of tumor necrosis factor-alpha and interleukin-10 (IL-10) were measured in serum. Lipid fractions, glucose and creatinine levels, and complete blood count were also assessed. Results: After controlling for a history of periodontitis, statistically significant differences between peri-implantitis patients and controls were found for total cholesterol (estimated adjusted mean difference, 76.4 mg/dL; 95% confidence interval [CI], 39.6, 113.2 mg/dL; P<0.001), low-density lipoprotein (LDL) cholesterol (estimated adjusted mean difference, 57.7 mg/dL; 95% CI, 23.8, 91.6 mg/dL; P<0.001), white blood cells (WBC) (estimated adjusted mean difference, 2.8×10³/µL; 95% CI, 1.6, 4.0×10³/µL; P<0.001) and IL-10 (estimated adjusted mean difference, -10.4 pg/mL; 95% CI, -15.8, -5.0 pg/mL; P<0.001). The peri-implant probing pocket depth (PPD) was modestly positively correlated with total cholesterol (r=0.512; P<0.001), LDL cholesterol (r=0.463; P=0.001), and WBC (r=0.519; P<0.001). A moderate negative correlation was observed between IL-10 and PPD (r=0.609; P<0.001). Conclusions: Otherwise healthy individuals with peri-implantitis showed increased low-grade systemic inflammation and dyslipidemia.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.17 no.3
• /
• pp.217-226
• /
• 2015

To investigate the distributions of BVOCs (Biogenic Volatile Organic Compounds) from mountain near mega city and their role in forest atmospheric, BVOCs and their oxidized species were measured at a 41 m tower in Mt. Taehwa during May, June and August 2013. A proton transfer reaction-mass spectrometer (PTR-MS) was used to quantify isoprene and monoterpenes. In conjunction with BVOCs, $O_3$, meteorological parameters, PAR (Photosynthetically Active Radiation) and LAI (Leaf Area Index) were measured. The average concentrations of isoprene and monoterpenes were 0.71 ppbv and 0.17 ppbv, respectively. BVOCs showed higher concentrations in the early summer (June) compared to the late summer (August). Isoprene started increasing at 2 PM and reached the maximum concentration around 5 PM. In contrast, monoterpenes concentrations began to increase 4 PM and stayed high at night. The $O_3$ maximum was generally found at 3 PM and remained high until 5 PM or later, which was concurrent with the enhancement of $O_3$. The concentrations of BVOCs were higher below canopy (18 m) than above canopy, which indicated these species were produced by trees. At night, monoterpenes concentrations were negatively correlated with these of $O_3$ below canopy. Using MEGAN (Model of Emissions of Gases and Aerosols from Nature), the emissions of isoprene and monoterpenes were estimated at 1.1 ton/year and 0.9 ton/year, respectively at Mt. Taehwa.

• The Mathematical Education
• /
• v.16 no.2
• /
• pp.22-26
• /
• 1978

It is shown that a map having an extension to an open map between the Alex-androff base compactifications of its domain and range has a unique such extension. J.S. Wasileski has introduced the Alexandroff base compactifications of Hausdorff spaces endowed with Alexandroff bases. We introduce a definition of morphism between such spaces to obtain a category which we denote by ABC. We prove that the Alexandroff base compactification on objects can be extended to a functor on ABC and that the compact objects give an epireflective subcategory of ABC. For each topological space X there exists a completely regular space $\alpha$X and a surjective continuous function $\alpha$$_{x}$ : Xlongrightarrow$\alpha$X such that for each completely regular space Z and g$\in$C (X, Z) there exists a unique g$\in$C($\alpha$X, 2) with g=g$^{\circ}$$\beta$$_{x}$. Such a pair ($\alpha$$_{x}$, $\alpha$X) is called a completely regularization of X. Let TOP be the category of topological spaces and continuous functions and let CREG be the category of completely regular spaces and continuous functions. The functor $\alpha$ : TOPlongrightarrowCREG is a completely regular reflection functor. For each topological space X there exists a compact Hausdorff space $\beta$X and a dense continuous function $\beta$x : Xlongrightarrow$\beta$X such that for each compact Hausdorff space K and g$\in$C (X, K) there exists a uniqueg$\in$C($\beta$X, K) with g=g$^{\circ}$$\beta$$_{x}$. Such a pair ($\beta$$_{x}$, $\beta$X) is called a Stone-Cech compactification of X. Let COMPT$_2$ be the category of compact Hausdorff spaces and continuous functions. The functor $\beta$ : TOPlongrightarrowCOMPT$_2$ is a compact reflection functor. For each topological space X there exists a realcompact space (equation omitted) and a dense continuous function (equation omitted) such that for each realcompact space Z and g$\in$C(X, 2) there exists a unique g$\in$C (equation omitted) with g=g$^{\circ}$(equation omitted). Such a pair (equation omitted) is called a Hewitt's realcompactification of X. Let RCOM be the category of realcompact spaces and continuous functions. The functor (equation omitted) : TOPlongrightarrowRCOM is a realcompact refection functor. In [2], D. Harris established the existence of a category of spaces and maps on which the Wallman compactification is an epirefiective functor. H. L. Bentley and S. A. Naimpally [1] generalized the result of Harris concerning the functorial properties of the Wallman compactification of a T$_1$-space. J. S. Wasileski [5] constructed a new compactification called Alexandroff base compactification. In order to fix our notations and for the sake of convenience. we begin with recalling reflection and Alexandroff base compactification.

• Sleep Medicine and Psychophysiology
• /
• v.3 no.2
• /
• pp.32-42
• /
• 1996

Objectives : Upper airway resistance syndrome(UARS) is a sleep-related breathing disorder characterized by abnormal negative intrathoracic pressure during sleep. Abnormally increased negative intrathoracic pressure results in microarousal and sleep fragmentation which underlay UARS-associated complaints of daytime fatigue and sleepiness. Although daytime dysfunction in patients with UARS is comparable to that of sleep apnea syndrome, UARS has been relatively unnoticed in clinical setting. That is why UARS is apt to be excluded in diagnosing of sleep-related breathing disorders since its respiratory disturbance index and arterial oxygen saturation are within normal limits. The current study presents a summary of clinical and polysomnographic characteristics found in patients with UARS. The present study aims (1) to explore characteristics of patients diagnosed with UARS, (2) to characterize the polysomnographic findings of UARS patients, and (3) to enhance the understanding of UARS through those clinical and laboratory characteristics. Methods : This was a retrospective study of 20 UARS patients (male 15, female 5) and 30 obstructive sleep apnea (OSA) patients (male 21, female 9) at the Stanford Sleep Disorders Clinic. We diagnosed patients as having UARS when they met critenia, RDI < 5 characteristic findings of an elevated esophageal pressure($<-10\;cmH_2O$), frequent arousals secondary to an elevated esophageal pressure, and symptoms of daytime fatigue and sleepiness. We used polysomnographic value, which is standardized by Williams et al(1974), as normal control. Statiotical test were done with student t-tests. Results : (1) Mean age of UARS was $41.0\;{\pm}\;14.8$ years and OSA was $50.9\;{\pm}\;12.0$ years. UARS subject was significantly younger than OSA subject (p<0.05). (2) The total score of Epworth Sleepiness Scale (ESS) was UARS $9.7\;{\pm}\;6.3$ and OSAS $11.2\;{\pm}\;6.3$. There was no significant difference between two groups. (3) The mean body mass index was UARS $28.1\;{\pm}\;5.7\;kg/m^2$ and OSAS $32.9\;{\pm}\;7.0\;kg/m^2$. UARS had significantly lower meen body man index than OSAS subjects (p<0.05). (4) The polysomnographic parameters of UARS were not significantly different from those of OSA except RDI(p<0.001), $SaO_2$ (p<0.001) and slow wave sleep latency (p<0.05). (5) Compared with normal control, Total sleep time in UARS subjects was significantly shorter (p<0.001), sleep efficiency index was significantly lower (p<0.001), total awakening percentage was significantly higher (p<0.001), and sleep stage 1 (p<0.001) were significantly higher. (6) OSA patients showed poor sleep quality and distinct abnormal sleep architectures compared with normal control. Conclusions : Conclusions from the above results are as follows : (1) UARS patients were younger and had lower body mass index when umpared with OSA patients. (2) The quality of sleep and sleep architectures of the UARS and OSA patients are significantly different from those of normal control. (3) ESS scores and awakening frequencies of UARS are similar with those of OSA, suggesting that daytime dysfunction of UARS patients may be comparable to those of OSA patients. (4) The RDI and the $SaO_2$ which are important indicators in diagnosing sleep-related breathing disorders, of UARS subjects are close to normal value. (5) According to the the above results, we unclude that despite the absence of $SaO_2$ drops and the absence of an elevated number of apnea and hypopnea, subjects developed clinical complaints which were associated with laborious breathing, elevated Pes nadir, and frequently snoring. (6) Accordingly, we suggest including LIARS in the differential diagnosis list when sleep related breathing disorder is suspected clinically and overnight polysomnographic findings except snoring and frequent microarousal are within normal limits.