• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.1-25
• /
• 2020

In this paper, we suggest an application system architecture which provides accurate, fast and efficient automatic gasometer reading function. The system captures gasometer image using mobile device camera, transmits the image to a cloud server on top of private LTE network, and analyzes the image to extract character information of device ID and gas usage amount by selective optical character recognition based on deep learning technology. In general, there are many types of character in an image and optical character recognition technology extracts all character information in an image. But some applications need to ignore non-of-interest types of character and only have to focus on some specific types of characters. For an example of the application, automatic gasometer reading system only need to extract device ID and gas usage amount character information from gasometer images to send bill to users. Non-of-interest character strings, such as device type, manufacturer, manufacturing date, specification and etc., are not valuable information to the application. Thus, the application have to analyze point of interest region and specific types of characters to extract valuable information only. We adopted CNN (Convolutional Neural Network) based object detection and CRNN (Convolutional Recurrent Neural Network) technology for selective optical character recognition which only analyze point of interest region for selective character information extraction. We build up 3 neural networks for the application system. The first is a convolutional neural network which detects point of interest region of gas usage amount and device ID information character strings, the second is another convolutional neural network which transforms spatial information of point of interest region to spatial sequential feature vectors, and the third is bi-directional long short term memory network which converts spatial sequential information to character strings using time-series analysis mapping from feature vectors to character strings. In this research, point of interest character strings are device ID and gas usage amount. Device ID consists of 12 arabic character strings and gas usage amount consists of 4 ~ 5 arabic character strings. All system components are implemented in Amazon Web Service Cloud with Intel Zeon E5-2686 v4 CPU and NVidia TESLA V100 GPU. The system architecture adopts master-lave processing structure for efficient and fast parallel processing coping with about 700,000 requests per day. Mobile device captures gasometer image and transmits to master process in AWS cloud. Master process runs on Intel Zeon CPU and pushes reading request from mobile device to an input queue with FIFO (First In First Out) structure. Slave process consists of 3 types of deep neural networks which conduct character recognition process and runs on NVidia GPU module. Slave process is always polling the input queue to get recognition request. If there are some requests from master process in the input queue, slave process converts the image in the input queue to device ID character string, gas usage amount character string and position information of the strings, returns the information to output queue, and switch to idle mode to poll the input queue. Master process gets final information form the output queue and delivers the information to the mobile device. We used total 27,120 gasometer images for training, validation and testing of 3 types of deep neural network. 22,985 images were used for training and validation, 4,135 images were used for testing. We randomly splitted 22,985 images with 8:2 ratio for training and validation respectively for each training epoch. 4,135 test image were categorized into 5 types (Normal, noise, reflex, scale and slant). Normal data is clean image data, noise means image with noise signal, relfex means image with light reflection in gasometer region, scale means images with small object size due to long-distance capturing and slant means images which is not horizontally flat. Final character string recognition accuracies for device ID and gas usage amount of normal data are 0.960 and 0.864 respectively.

• Journal of the Korean Society of Radiology
• /
• v.8 no.2
• /
• pp.89-96
• /
• 2014

This paper tries to examine whether the application of joint mobilization to subjects who have the forward head posture due to malalignment in the cervical joint has influence on posture changes and functions in the cervical joint. The subjects were 39 students from G University in Gyeongsangbuk-do. The cervical joint mobilization was applied to 20 subjects and not to 19. The students with a cervical lordosis angle of $21^{\circ}C$ or less, an anterior weight bearing (AWB) of 15mm or greater, and a cervical extension ROM of $70^{\circ}C$ or less in terms of radiography were selected as subjects under their voluntary agreement. The patients actively performed the joint mobilization slowly 8 times per session while therapists continuously applied sustained accessory glide to their painful joints 3 times per week for 4 weeks along with the cervical expansion and flexion in SNAGS among other Mulligan's (1995) techniques. The measurement was carried out in terms of radiographic inspection and neck disability indexes. As a result of the experiment, it turned out that the subjects with the forward head posture had changes in the cervical AWB and ARA, the ranges of expansion and flexion, and the NDI(Neck Disability Index) after the intervention for the experimental group by applying cervical joint mobilization. There were no changes observed in the control group. In conclusion, the application of joint mobilization turned out to have influence on the improvement of cervical joint postures, and craniocervical region functions.

• Tuberculosis and Respiratory Diseases
• /
• v.42 no.5
• /
• pp.654-659
• /
• 1995

Background: Tuberculosis has commonly been associated with various hematologic changes. A difference between the changes found in pulmonary tuberculosis and those found in miliary tuberculosis has been discussed. Up to now some worker were investigated hematological changes associated with pulmonary tuberculosis but was not investigated those associated with miliary tuberculosis in korea. Therefore we examimed the peripheral blood and bone marrow findings in miliary tuberculosis patients to determine hematologic changes. Methods: We performed blood sample at admission and bone marrow biopsy within 7days after admission. For evaluation of the hematologic findings, full blood counts and marrow differential counts were defined by the criteria outlined by Dacie and Lewis. Results: 1) Peripheral blood findings: Pancytopenia in 10% of patients, anemia in all patients, leukocytosis in 10% of patients, leukopenia in 20% of patients, thrombocytopenia in 30% of patients, lymphocytopenia in 90% of patients, monocytosis in 40% of patients and neutrophilia in 10% of patients were found at peripheral blood. 2) Bone marrow findings: Lymphocytopenia in 30% of patients, lymphocytosis in 20% of patients, plasmacytosis in 40% of patients, monocytosis in 100% of patients, and hypocellularity in 30% of patients were found at bone marrow. Erythropoiesis was decreased in 30% of patients. Granulopoiesis was decreased in 20% of patients and increased in 20% of patients. Bone morrow granuloma occured in 25% of patients. Conclusion: Hematologic changes of miliary tuberculosis were seen tendency of cytopenic pattern but monocyte was increased at peripheral blood and bone morrow. This findings would provide additional information for the differential diagnosis of miliary tuberculosis.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.25 no.4
• /
• pp.327-336
• /
• 2007

The 3D building modeling is one of crucial components in building 3D geospatial information. The existing methods for 3D building modeling depend mainly on manual photogrammetric processes by stereoplotter compiler, which indeed take great amount of time and efforts. In addition, some automatic methods that were proposed in research papers and experimental trials have limitations of describing the details of buildings with lack of geometric accuracy. It is essential in automatic fashion that the boundary and shape of buildings should be drawn effortlessly by a sophisticated algorithm. In recent years, airborne LiDAR data representing earth surface in 3D has been utilized in many different fields. However, it is still in technical difficulties for clean and correct boundary extraction without human intervention. The usage of airborne LiDAR data will be much feasible to reconstruct the roof tops of buildings whose boundary lines could be taken out from existing digital maps. The paper proposed a method to reconstruct the roof tops of buildings using airborne LiDAR data with building boundary lines from digital map. The primary process is to perform octree-based segmentation to airborne LiDAR data recursively in 3D space till there are no more airborne LiDAR points to be segmented. Once the octree-based segmentation has been completed, each segmented patch is thereafter merged based on geometric spatial characteristics. The experimental results showed that the proposed method were capable of extracting various building roof components such as plane, gable, polyhedric and curved surface.

• Korean Journal of Social Welfare Studies
• /
• v.48 no.2
• /
• pp.5-33
• /
• 2017

This study explored challenges and desires related to social relationships among adolescents with intellectual disabilities, using the photovoice method. Six adolescents with intellectual disabilities (M = 18.33 years) who are attending a general high school participated in 7 sessions of a photovoice program. They selected themes related to social relationships through group discussions and shared their photos related to the themes with other participants. Our study yielded 15 subthemes within 5 social relationships (peers, teachers, boyfriends/girlfriends, parents, persons whom I need) from adolescents with intellectual disabilities as follows: relationships with peers (feeling lonely and left out, being teased and bullied by peers, making efforts to fix damaged relationships with friends), relationships with teachers (feeling appreciated by teachers who understand me, lacking interactions with teachers, wanting teachers to show they care about my school life), relationships with boyfriends/girlfriends (wanting to have a boyfriend/girlfriend, feeling worried and fearful about having a boyfriend/girlfriend, trying to attract a boyfriend/girlfriend), relationships with parents (being thankful to parents for their unconditioned love, enjoying time with parents, feeling bad because of not being close to father), persons whom I need (persons who enjoy spending time with me, persons who help me, persons who never give up on me). This study provides implications for social work practice, such as training for improving peer relationships and coping with peer conflict, education for fostering healthy romantic relationships, training for promoting social skills of adolescents with intellectual disabilities, and education for increasing awareness of the social needs of students with intellectual disabilities among peers without disabilities, teachers, and social workers.

• Cross-Cultural Studies
• /
• v.49
• /
• pp.59-93
• /
• 2017

This paper provides an in-depth examination of the relationship between cinema and Giorgio Agamben's aesthetics and philosophy. Intersecting Agamben's key concepts including gesture, mediality, biopolitics, historicity, and profanation with historical and aesthetic dimensions of cinema, I argue for his ambivalent view on cinema and visual media. On the one hand, Agamben linked cinema and visual media to his discussion on biopolitics and spectacle as he considered them as apparatus for capturing and controlling gestures. On the other hand, he also argued that cinema could restore the image with capacity to preserve and recuperate gestures based on his consideration of montage as cinema's key aesthetic and technical component (an operation of profanation) and his Benjaminian thought on the ways in which montage suspended linear flow of images and activated an alternative memory of them. Drawing on history of cinema and optical devices in the 19th and early 20th centuries as well as examples of found footages of filmmaking predicated upon stoppage and repetition of images, I argue that Agamben's concept of potentialities can be extended into his thought on cinema and visual media apparatuses in general.

• Journal of Dental Rehabilitation and Applied Science
• /
• v.38 no.4
• /
• pp.196-203
• /
• 2022

Purpose: This study aimed to evaluate the accuracy and clinical usability of an identification model using deep learning for 79 dental implant types. Materials and Methods: A total of 45396 implant fixture images were collected through panoramic radiographs of patients who received implant treatment from 2001 to 2020 at 30 dental clinics. The collected implant images were 79 types from 18 manufacturers. EfficientNet and Meta Pseudo Labels algorithms were used. For EfficientNet, EfficientNet-B0 and EfficientNet-B4 were used as submodels. For Meta Pseudo Labels, two models were applied according to the widen factor. Top 1 accuracy was measured for EfficientNet and top 1 and top 5 accuracy for Meta Pseudo Labels were measured. Results: EfficientNet-B0 and EfficientNet-B4 showed top 1 accuracy of 89.4. Meta Pseudo Labels 1 showed top 1 accuracy of 87.96, and Meta pseudo labels 2 with increased widen factor showed 88.35. In Top5 Accuracy, the score of Meta Pseudo Labels 1 was 97.90, which was 0.11% higher than 97.79 of Meta Pseudo Labels 2. Conclusion: All four deep learning algorithms used for implant identification in this study showed close to 90% accuracy. In order to increase the clinical applicability of deep learning for implant identification, it will be necessary to collect a wider amount of data and develop a fine-tuned algorithm for implant identification.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.50 no.6
• /
• pp.110-123
• /
• 2022

Namsan Park in Seoul was designated as a "grand park" in 1954 and is currently operated as an 'Urban Nature Park Area' and four 'neighborhood parks.' However, despite the park's historical and cultural value as an urban park, it has been discussed mainly from a perspective revolving around notions of a mountain or a city wall. To ensure a comprehensive exploration of Namsan Park's history, this study examined public records at the Seoul Metropolitan Archives (SMA), which houses the city's permanent records for preservation and organization. To this end, documents in the SMA Database (DB) were analyzed, yielding 1,359 records concerning Namsan Park. Based on the contents, general characteristics of the urban park were identified through production periods, record types, and disclosure types. Then, essential keywords concerning organizations, people, geographical areas, subjects, and business functions were examined. Finally, the contents and characteristics of Namsan Park in public records were scrutinized, focusing on specific spaces. This research also uncovered important information, such as park drawings, photos, planting lists, plant parcel lists, and significant discussions and decisions regarding the operation and management of the park. Although the public records do not contain a comprehensive history of Namsan Park, it was possible to check the primary historical changes and deliberation processes pertaining to the park's history. Therefore, continuous research intended to interpret and describe public records is expected to identify many implications. In addition, because the public records showed heterogeneous characteristics that center on specific periods and events, an essential task is to advance collaboration and networking with various related institutions, designers, researchers, and citizens.

• Journal of the korean academy of Pediatric Dentistry
• /
• v.31 no.1
• /
• pp.66-78
• /
• 2004

The objective of this study was to apply the vibration technique to reduce the viscosity of bonding adhesives and thereby compare the bond strength and resin penetration in enamel and dentin achieved with those gained using the conventional technique and vibration technique. For enamel specimens, thirty teeth were sectioned mesio-distally. Sectioned two parts were assigned to same adhesive system but different treatment(vibration vs. non-vibration). Each specimen was embedded in 1-inch inner diameter PVC pipe with a acrylic resin. The buccal and lingual surfaces were placed so that the tooth and the embedding medium were at the same level. The samples were subsequently polished silicon carbide abrasive papers. Each adhesive system was applied according to its manufacture's instruction. Vibration groups were additionally vibrated for 15 seconds before curing. For dentin specimen, except removing the coronal part and placing occlusal surface at the mold level, the remaining procedures were same as enamel specimen. Resin composite(Z250. 3M. U.S.A.) was condensed on to the prepared surface in two increments using a mold kit(Ultradent Inc., U.S.A.). Each increments was light cured for 40 seconds. After 24 hours in tap water at room temperature, the specimens were thermocycled for 1000cycles. Shear bond strengths were measured with a universal testing machine(Instron 4465, England). To investigate infiltration patterns of adhesive materials, the surface of specimens was examined with scanning electron microscope. The results were as follows: 1. In enamel the mean values of shear bond strengths in vibration groups(group 2, 4, 6) were greater than those of non-vibration group(group 1, 3, 5). The differences were statistically significant except AQ bond group. 2. In dentin, the mean values of shear bond strengths in vibration groups(group 2, 4, 6) were greater than those of non-vibration groups(group 1, 3, 5). But the differences were not statistically significant except One-Up Bond F group. 3. The vibration group showed more mineral loss in enamel and longer resin tag and greater number of lateral branches in dentin under SEM examination.

• Journal of the Korea Computer Graphics Society
• /
• v.6 no.4
• /
• pp.1-7
• /
• 2000

Traditionally, three dimension model s have been used for building virtual worlds, and a data structure called the "scene graph" is often employed to organize these 3D objects in the virtual space. On the other hand, image-based rendering has recently been suggested as a probable alternative VR platform for its photo-realism, however, due to limited interactivity. it has only been used for simple navigation systems. To combine the merits of these two approaches to object/scene representations, this paper proposes for a scene graph structure in which both 3D models and various image-based scenes/objects can be defined. traversed, and rendered together. In fact, as suggested by Shade et al. [1]. these different representations can be used as different LOD's for a given object. For in stance, an object might be rendered using a 3D model at close range, a billboard at an intermediate range. and as part of an environment map at far range. The ultimate objective of this mixed platform is to breath more interactivity into the image based rendered VE's by employing 3D models as well. There are several technical challenges in devising such a platform : designing scene graph nodes for various types of image based techniques, establishing criteria for LOD/representation selection. handling their transition s. implementing appropriate interaction schemes. and correctly rendering the overall scene. Currently, we have extended the scene graph structure of the Sense8's WorldToolKit. to accommodate new node types for environment maps. billboards, moving textures and sprites, "Tour-into-the-Picture" structure, and view interpolated objects. As for choosing the right LOD level, the usual viewing distance and image space criteria are used, however, the switching between the image and 3D model occurs at a distance from the user where the user starts to perceive the object's internal depth. Also. during interaction, regardless of the viewing distance. a 3D representation would be used, if it exists. Finally. we carried out experiments to verify the theoretical derivation of the switching rule and obtained positive results.