• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.9
• /
• pp.371-380
• /
• 2022

Convolutional neural networks are widely used to manipulate data arranged in a grid, such as images. A general convolutional neural network consists of a convolutional layers and a fully connected layers, and each layer contains a nonlinear activation functions. This paper proposes a combined parametric activation function to improve the performance of convolutional neural networks. The combined parametric activation function is created by adding the parametric activation functions to which parameters that convert the scale and location of the activation function are applied. Various nonlinear intervals can be created according to parameters that convert multiple scales and locations, and parameters can be learned in the direction of minimizing the loss function calculated by the given input data. As a result of testing the performance of the convolutional neural network using the combined parametric activation function on the MNIST, Fashion MNIST, CIFAR10 and CIFAR100 classification problems, it was confirmed that it had better performance than other activation functions.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.8
• /
• pp.347-354
• /
• 2022

Since the era of the 4th industrial revolution has been ripe, the use of artificial intelligence(AI) based on massive data is beginning to be actively applied in various fields. However, as the process of analyzing forest species is carried out manually, many errors are occurring. Therefore, in this paper, about 60,000 pieces of AI learning data were automatically analyzed for pine, larch, conifer, and broadleaf trees of aerial photographs and pseudo images in the metropolitan area, and an AI model was developed to distinguish tree species. Through this, it is expected to increase in work efficiency by using the tree species division image as basic data when producing forest change detection and forest field topics.

• Journal of Microbiology and Biotechnology
• /
• v.32 no.12
• /
• pp.1589-1598
• /
• 2022

Microfibrillated cellulose (MFC) is a valuable material with wide industrial applications, particularly for the food and cosmetics industries, owing to its excellent physiochemical properties. Here, we prepared high-solid microfibrillated cellulose (HMFC) from the centrifugation of Gelidium amansiiderived MFC right after fibrillation. Dispersion properties, morphology, and structural changes were monitored during processing. HMFC has a five-fold higher solid concentration than MFC without significant changes to dispersion properties. SEM images and FTIR spectra of HMFC revealed a stable surface and structure against centrifugal forces. HMFC exhibited 2,2'-azino-bis (3-ethylbenzothiazoline6-sulfonic acid) (ABTS) radical scavenging activity, although it could not scavenge 2,2-diphenyl-1- picrylhydrazyl (DPPH). Moreover, HMFC inhibited the generation of LPS-induced excessive nitrite and radial oxygen species in murine macrophage RAW264.7 cells. Additionally, HMFC suppressed LPS-induced Keap-1 expression in the cytosol but did not alter iNOS expression. HMFC also attenuated the UVB-induced phosphorylation of p38, c-Jun N-terminal kinase (JNK) 1/2, and extracellular-signal-regulated kinase (ERK) 1/2, as well as the phosphorylation of c-Jun in the immortalized human skin keratinocyte HaCaT cells. Therefore, the application of centrifugation is suitable for producing high-solid MFC as a candidate material for anti-inflammatory and antioxidative marine cosmeceuticals.

• Imaging Science in Dentistry
• /
• v.52 no.4
• /
• pp.383-391
• /
• 2022

Purpose: Periodontitis, the most prevalent chronic inflammatory condition affecting teeth-supporting tissues, is diagnosed and classified through clinical and radiographic examinations. The staging of periodontitis using panoramic radiographs provides information for designing computer-assisted diagnostic systems. Performing image segmentation in periodontitis is required for image processing in diagnostic applications. This study evaluated image segmentation for periodontitis staging based on deep learning approaches. Materials and Methods: Multi-Label U-Net and Mask R-CNN models were compared for image segmentation to detect periodontitis using 100 digital panoramic radiographs. Normal conditions and 4 stages of periodontitis were annotated on these panoramic radiographs. A total of 1100 original and augmented images were then randomly divided into a training (75%) dataset to produce segmentation models and a testing (25%) dataset to determine the evaluation metrics of the segmentation models. Results: The performance of the segmentation models against the radiographic diagnosis of periodontitis conducted by a dentist was described by evaluation metrics(i.e., dice coefficient and intersection-over-union [IoU] score). MultiLabel U-Net achieved a dice coefficient of 0.96 and an IoU score of 0.97. Meanwhile, Mask R-CNN attained a dice coefficient of 0.87 and an IoU score of 0.74. U-Net showed the characteristic of semantic segmentation, and Mask R-CNN performed instance segmentation with accuracy, precision, recall, and F1-score values of 95%, 85.6%, 88.2%, and 86.6%, respectively. Conclusion: Multi-Label U-Net produced superior image segmentation to that of Mask R-CNN. The authors recommend integrating it with other techniques to develop hybrid models for automatic periodontitis detection.

• Journal of radiological science and technology
• /
• v.45 no.6
• /
• pp.553-560
• /
• 2022

The purpose of this study is to evaluate the clinical risk according to the applicator heterogeneity, mislocation, and tissue heterogeneity correction through a dose verification program during brachytherapy of cervical cancer. We performed image processing with MATLAB on images acquired with CT simulator. The source was modeled and stochiometric calibration and Monte-Carlo algorithm were applied based on dwell time and location to calculate the dose, and the secondary cancer risk was evaluated in the dose verification program. The result calculated by correcting for applicator and tissue heterogeneity showed a maximum dose of about 25% higher. In the bladder, the difference in excess absolute risk according to the heterogeneity correction was not significant. In the rectum, the difference in excess absolute risk was lower than that calculated by correcting applicator and tissue heterogeneity compared to the water-based calculation. In the femur, the water-based calculation result was the lowest, and the result calculated by correcting the applicator and tissue heterogeneity was 10% higher. A maximum of 14% dose difference occurred when the applicator mislocation was 20 mm in the Z-axis. In a future study, it is expected that a system that can independently verify the treatment plan can be developed by automating the interface between the treatment planning system and the dose verification program.

• Journal of the Society of Naval Architects of Korea
• /
• v.60 no.2
• /
• pp.120-134
• /
• 2023

In this study, hydrodynamic characteristics of dam break flow were investigated by a series of experiments. The experiments were performed in a 2-D rectangular flume with obtaining instantaneous images of dam break flow to capture the free surface elevation, and pressure distributions on vertical wall and bottom of the flume. The initial water depth of the dam break flow was changed into 3 different heights, and the gate opening speed was changed during the experiments to study the effect of the gate speed in the dam break flow. Generation of dam break phenomena could be classified into three stages, i.e., very initial, relatively stable, and wall impact stages. The wall impact stage could be separated into 4 generation phases of wall impinge, run-up, overturning, and touchdown phases based on the deformation of the free surface. The free surface elevation were investigated with various initial water depth and compared with the analytic solutions by Ritter (1892). The pressures acting on the vertical wall and bottom were provided for the whole period of dam break flow varying the initial water depth and gate open speed. The measurement results of the dam break flow was compared with the hydrodynamic characteristics of green water phenomena, and it showed that the dam break flow could overestimate the green water loading based on the estimation suggested by Buchner (2002).

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.535-538
• /
• 2022

For Plug & Play of IoT devices, we develop a module that recognizes the type of USB, which is a typical wired interface of IoT devices, through image recognition. In order to drive an IoT device, a driver for communication and device hardware is required. The wired interface for connecting to the IoT device is recognized by using the image obtained through the camera of smartphone shooting to recognize the corresponding communication interface. For USB, which is a most popular wired interface, types of USB are classified through artificial neural network-based machine learning. In order to secure sufficient data set of artificial neural networks, USB images are collected through the Internet, and additional image data sets are secured through image processing. In addition to the convolution neural networks, recognizers are implemented with various deep artificial neural networks, and their performance is compared and evaluated.

• Journal of the Korean Geotechnical Society
• /
• v.23 no.12
• /
• pp.109-116
• /
• 2007

This study has reviewed the applicability of displacement measurement by using a digital vision technique based on typical photogrammetric methods. In this study, a series of experimental measurements have been performed in order to improve the accuracy of digital vision measurement by establishing criteria of factors of various vision measurements. It is found that the digital vision measurement tends to show higher accuracy as the image size(resolution) and the focal length become larger and the distance to an object becomes closer. It is also observed that measurement error decreases with processing as many images as possible in various angles. Applicability on high-resolution displacement measurement is proved by applying the digital vision measurement developed in this study to a large scale loading test of concrete lining.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.12
• /
• pp.509-516
• /
• 2022

Accurate pose prediction of objects in 3D space is an important visual recognition technique widely used in many applications such as scene understanding in both indoor and outdoor environments, robotic object manipulation, autonomous driving, and augmented reality. Most previous works for object pose estimation have the limitation that they require an exact 3D CAD model for each object. Unlike such previous works, this paper proposes a novel neural network model that can predict the poses of unknown objects based on only their RGB color images without the corresponding 3D CAD models. The proposed model can obtain depth maps required for unknown object pose prediction by using an adaptive depth estimator, AdaBins,. In this paper, we evaluate the usefulness and the performance of the proposed model through experiments using benchmark datasets.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.3
• /
• pp.203-213
• /
• 2023

In medical ultrasound imaging systems, Side lobes may appear if signals outside the imaging point are not completely removed during receive focusing. If the time signal of the side lobe overlaps with the time signal (main lobe) from the image point, it is difficult to completely remove it using filter processing in the time domain. However, In the receive focusing process, when time-channel signals are Fourier-transformed, the main lobe and side lobe signals are spatially separated in the spectral domain. Therefore, the side lobes can be suppressed by multiplying the image with magnitude weights, which are determined by the magnitudes of the main and side lobes calculated in the spectral domain. In addition, when the main lobe and the side lobe spectrum are adjacent, the distance weight was applied based on the distance between them. In a 5 MHz ultrasound imaging system using a 64-channel linear transducer, point reflector and speckle images with cysts of various brightness were synthesized and weights were applied to the ultrasound image. Using computer simulations, we confirmed that the side lobes were greatly reduced without affecting the spatial resolution in the point reflector image, and the contrast was significantly improved in the cyst image with computer simulations.