• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.6
• /
• pp.267-275
• /
• 2023

This paper presents the development of specialized software for annotating volume-of-interest on ¹⁸F-FDG PET/CT images with the goal of facilitating the studies and diagnosis of head and neck cancer (HNC). To achieve an efficient annotation process, we employed the SE-Norm-Residual Layer-based U-Net model. This model exhibited outstanding proficiency to segment cancerous regions within ¹⁸F-FDG PET/CT scans of HNC cases. Manual annotation function was also integrated, allowing researchers and clinicians to validate and refine annotations based on dataset characteristics. Workspace has a display with fusion of both PET and CT images, providing enhance user convenience through simultaneous visualization. The performance of deeplearning model was validated using a Hecktor 2021 dataset, and subsequently developed semi-automatic annotation functionalities. We began by performing image preprocessing including resampling, normalization, and co-registration, followed by an evaluation of the deep learning model performance. This model was integrated into the software, serving as an initial automatic segmentation step. Users can manually refine pre-segmented regions to correct false positives and false negatives. Annotation images are subsequently saved along with their corresponding ¹⁸F-FDG PET/CT fusion images, enabling their application across various domains. In this study, we developed a semi-automatic annotation software designed for efficiently generating annotated lesion images, with applications in HNC research and diagnosis. The findings indicated that this software surpasses conventional tools, particularly in the context of HNC-specific annotation with ¹⁸F-FDG PET/CT data. Consequently, developed software offers a robust solution for producing annotated datasets, driving advances in the studies and diagnosis of HNC.

• Journal of IKEEE
• /
• v.28 no.2
• /
• pp.168-173
• /
• 2024

Speaker recognition refers to a technology that analyzes voice frequencies that are different for each individual and compares them with pre-stored voices to determine the identity of the person. Deep learning-based speaker recognition is being applied to many fields, and pet robots are one of them. However, the hardware performance of pet robots is very limited in terms of the large memory space and calculations of deep learning technology. This is an important problem that pet robots must solve in real-time interaction with users. Lightening deep learning models has become an important way to solve the above problems, and a lot of research is being done recently. In this paper, we describe the results of research on lightweight speaker recognition for pet robots by constructing a voice data set for pet robots, which is a specific command type, and comparing the results of models using residuals. In the conclusion, we present the results of the proposed method and Future research plans are described.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.28 no.4
• /
• pp.193-203
• /
• 2024

Existing reinforced concrete (RC) building frames constructed before the seismic design was applied have seismically deficient structural details, and buildings with such structural details show brittle behavior that is destroyed early due to low shear performance. Various reinforcement systems, such as fiber-reinforced polymer (FRP) jacketing systems, are being studied to reinforce the seismically deficient RC frames. Due to the step-by-step modeling and interpretation process, existing seismic performance assessment and reinforcement design of buildings consume an enormous amount of workforce and time. Various machine learning (ML) models were developed using input and output datasets for seismic loads and reinforcement details built through the finite element (FE) model developed in previous studies to overcome these shortcomings. To assess the performance of the seismic performance prediction models developed in this study, the mean squared error (MSE), R-square (R²), and residual of each model were compared. Overall, the applied ML was found to rapidly and effectively predict the seismic performance of buildings according to changes in load and reinforcement details without overfitting. In addition, the best-fit model for each seismic performance class was selected by analyzing the performance by class of the ML models.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.5 no.3
• /
• pp.141-145
• /
• 2009

In this paper, we propose an efficient lossless compression algorithm using spatial and temporal information. The proposed method obtains higher lossless compression of images than other lossless compression techniques. It is divided into two parts, a motion adaptation based predictor part and a residual error coding part. The proposed nonlinear predictor can reduce prediction error by learning from its past prediction errors. The predictor decides the proper selection of the spatial and temporal prediction values according to each past prediction error. The reduced error is coded by existing context coding method. Experimental results show that the proposed algorithm has better performance than those of existing context modeling methods.

• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10b
• /
• pp.1-6
• /
• 1992

A principle of "orthogonalization" is proposed as an extended notion of hybrid (force and position) control for robot manipulators under geometric endpoint constraints. The principle realizes the hybrid control in a strict sense by letting position and velocity feedback signals be orthogonal in joint space to the contact force vector whose components are exerted at corresponding joints. This orthogonalization is executed via a projection matrix computed in real-time from a gradient of the equation of the surface in joint coordinates and hence both projected position and velocity feedback signals become perpendicular to the force vector that is normal to the surface at the contact point in joint space. To show the important role of the principle in control of robot manipulators, three basic problems are analyzed, the first is a hybrid trajectory tracking problem by means of a "modified hybrid computed torque method", the second is a model-based adaptive control problem for robot manipulators under geometric endpoint constraints, and the third is an iterative learning control problem. It is shown that the passivity of residual error dynamics of robots follows from the orthogonalization principle and it plays a crucial role in convergence properties of both positional and force error signals.force error signals.

• Sleep Medicine and Psychophysiology
• /
• v.12 no.2
• /
• pp.98-104
• /
• 2005

Approximately 1% to 3% of all children have obstructive sleep apnea syndrome (OSAS). OSAS in children can lead to a variety of symptoms and sequalae; impairment of development and quality of life, behavioral and personality disturbance, learning problem, cor pulmonale and hypertension. Diagnosis and treatment of OASA for children are different from those for adults in many respects. Adenotonsillar hypertrophy is major cause of childhood OSAS. Overnight polysomnography in a sleep laboratory is the gold standard for diagnosing childhood OSAS. However, because full polysomnography in children may be difficult to obtain, expensive, and inconvenient, other methods to diagnose OSAS have been investigated. Adenotonsillectomy is the most common surgical treatment of childhood OSAS. But if residual symptoms remained after adenotonsillectomy, it should be considered to additional treatment such as weight control, sleep positional change, and continuous positive airway pressure (CPAP).

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.9-14
• /
• 2002

We propose a new algorithm for blind source separation (BSS), in which frequency-domain independent component analysis (FDICA) and time-domain ICA (TDICA) are combined to achieve a superior source-separation performance under reverberant conditions. Generally speaking, conventional TDICA fails to separate source signals under heavily reverberant conditions because of the low convergence in the iterative learning of the inverse of the mixing system. On the other hand, the separation performance of conventional FDICA also degrades significantly because the independence assumption of narrow-band signals collapses when the number of subbands increases. In the proposed method, the separated signals of FDICA are regarded as the input signals for TDICA, and we can remove the residual crosstalk components of FDICA by using TDICA. The experimental results obtained under the reverberant condition reveal that the separation performance of the proposed method is superior to that of conventional ICA-based BSS methods.

• Korean Journal of Cleft Lip And Palate
• /
• v.11 no.2
• /
• pp.77-82
• /
• 2008

The harmonious lip length, residual scar and Cupid's bow width and peak with a normal side are the aims of a unilateral cleft lip repair. Also, long term amelioration without necessary of revisional surgeries may be the ideal conditions. No one method can satisfy the wide varieties of cleft lip deformities. Recently with rearrangement of paraoral muscle and some modifications, a straight line repair technique has been concerned again. Straight scar line, simplicity, and short learning curve are the advantages of the straight line technique. Here two cases of the simple straight line technique were presented and discussed for its usefulness and reliability with short reviews of previous reports.

• Journal of Korea Multimedia Society
• /
• v.25 no.11
• /
• pp.1582-1590
• /
• 2022

Due to the situation of the widespread of the coronavirus, which causes the problem of lack of face image data occluded by masks at recent time, in order to solve the related problems, this paper proposes a method to generate face images with masks using a combination of generative adversarial networks and spatial transformation networks based on CNN model. The system we proposed in this paper is based on the GAN, combined with multi-scale convolution kernels to extract features at different details of the human face images, and used Wasserstein divergence as the measure of the distance between real samples and synthetic samples in order to optimize Generator performance. Experiments show that the proposed method can effectively put masks on face images with high efficiency and fast reaction time and the synthesized human face images are pretty natural and real.

• Archives of Craniofacial Surgery
• /
• v.22 no.5
• /
• pp.223-231
• /
• 2021

The field of artificial intelligence (AI) is rapidly advancing, and AI models are increasingly applied in the medical field, especially in medical imaging, pathology, natural language processing, and biosignal analysis. On the basis of these advances, telemedicine, which allows people to receive medical services outside of hospitals or clinics, is also developing in many countries. The mechanisms of deep learning used in medical AI include convolutional neural networks, residual neural networks, and generative adversarial networks. Herein, we investigate the possibility of using these AI methods in the field of craniofacial surgery, with potential applications including craniofacial trauma, congenital anomalies, and cosmetic surgery.