• The Journal of Korean Institute of Information Technology
• /
• v.17 no.6
• /
• pp.17-25
• /
• 2019

Accurate tracking of the sinkholes that are appearing frequently now is an important method of protecting human and property damage. Although many sinkhole detection systems have been proposed, it is still far from completely solved especially in-depth area. Furthermore, detection of sinkhole algorithms experienced the problem of unstable result that makes the system difficult to fire a warning in real-time. In this paper, we proposed a method of sinkhole tracking by deep learning and data association, that takes advantage of the recent development of CNN transfer learning. Our system consists of three main parts which are binary segmentation, sinkhole classification, and sinkhole tracking. The experiment results show that the sinkhole can be tracked in real-time on the dataset. These achievements have proven that the proposed system is able to apply to the practical application.

• Clinical Endoscopy
• /
• v.55 no.3
• /
• pp.390-400
• /
• 2022

Background/Aims: Previous artificial intelligence (AI) models attempting to segment gastric intestinal metaplasia (GIM) areas have failed to be deployed in real-time endoscopy due to their slow inference speeds. Here, we propose a new GIM segmentation AI model with inference speeds faster than 25 frames per second that maintains a high level of accuracy. Methods: Investigators from Chulalongkorn University obtained 802 histological-proven GIM images for AI model training. Four strategies were proposed to improve the model accuracy. First, transfer learning was employed to the public colon datasets. Second, an image preprocessing technique contrast-limited adaptive histogram equalization was employed to produce clearer GIM areas. Third, data augmentation was applied for a more robust model. Lastly, the bilateral segmentation network model was applied to segment GIM areas in real time. The results were analyzed using different validity values. Results: From the internal test, our AI model achieved an inference speed of 31.53 frames per second. GIM detection showed sensitivity, specificity, positive predictive, negative predictive, accuracy, and mean intersection over union in GIM segmentation values of 93%, 80%, 82%, 92%, 87%, and 57%, respectively. Conclusions: The bilateral segmentation network combined with transfer learning, contrast-limited adaptive histogram equalization, and data augmentation can provide high sensitivity and good accuracy for GIM detection and segmentation.

• Journal of Internet Computing and Services
• /
• v.23 no.4
• /
• pp.35-43
• /
• 2022

Makeup is the most common way to improve a person's appearance. However, since makeup styles are very diverse, there are many time and cost problems for an individual to apply makeup directly to himself/herself.. Accordingly, the need for makeup automation is increasing. Makeup transfer is being studied for makeup automation. Makeup transfer is a field of applying makeup style to a face image without makeup. Makeup transfer can be divided into a traditional image processing-based method and a deep learning-based method. In particular, in deep learning-based methods, many studies based on Generative Adversarial Networks have been performed. However, both methods have disadvantages in that the resulting image is unnatural, the result of makeup conversion is not clear, and it is smeared or heavily influenced by the makeup style face image. In order to express the clear boundary of makeup and to alleviate the influence of makeup style facial images, this study divides the makeup area and calculates the loss function using HoG (Histogram of Gradient). HoG is a method of extracting image features through the size and directionality of edges present in the image. Through this, we propose a makeup transfer network that performs robust learning on edges.By comparing the image generated through the proposed model with the image generated through BeautyGAN used as the base model, it was confirmed that the performance of the model proposed in this study was superior, and the method of using facial information that can be additionally presented as a future study.

• Computers and Concrete
• /
• v.30 no.1
• /
• pp.43-74
• /
• 2022

Machine learning technique is recently opening new opportunities to identify the complex shear transfer mechanisms of reinforced concrete (RC) beam members. This study employed 1224 shear test specimens to train decision tree-based machine learning (ML) programs, by which strong correlations between shear capacity of RC beams and key input parameters were affirmed. In addition, shear contributions of concrete and shear reinforcement (the so-called V_c and V_s) were identified by establishing three independent ML models trained under different strategies with various combinations of datasets. Detailed parametric studies were then conducted by utilizing the well-trained ML models. It appeared that the presence of shear reinforcement can make the predicted shear contribution from concrete in RC beams larger than the pure shear contribution of concrete due to the intervention effect between shear reinforcement and concrete. On the other hand, the size effect also brought a significant impact on the shear contribution of concrete (V_c), whereas, the addition of shear reinforcements can effectively mitigate the size effect. It was also found that concrete tends to be the primary source of shear resistance when shear span-depth ratio a/d<1.0 while shear reinforcements become the primary source of shear resistance when a/d>2.0.

• International Journal of Computer Science & Network Security
• /
• v.22 no.10
• /
• pp.257-261
• /
• 2022

Diseases of agricultural plants in recent years have spread greatly across the regions of the Kyrgyz Republic and pose a serious threat to the yield of many crops. The consequences of it can greatly affect the food security for an entire country. Due to force majeure, abnormal cases in climatic conditions, the annual incomes of many farmers and agricultural producers can be destroyed locally. Along with this, the rapid detection of plant diseases also remains difficult in many parts of the regions due to the lack of necessary infrastructure. In this case, it is possible to pave the way for the diagnosis of diseases with the help of the latest achievements due to the possibilities of feedback from the farmer - developer in the formation and updating of the database of sick and healthy plants with the help of advances in computer vision, developing on the basis of machine and deep learning. Currently, model training is increasingly used already on publicly available datasets, i.e. it has become popular to build new models already on trained models. The latter is called as transfer training and is developing very quickly. Using a publicly available data set from PlantVillage, which consists of 54,306 or NewPlantVillage with a data volumed with 87,356 images of sick and healthy plant leaves collected under controlled conditions, it is possible to build a deep convolutional neural network to identify 14 types of crops and 26 diseases. At the same time, the trained model can achieve an accuracy of more than 99% on a specially selected test set.

• The Bulletin of The Korean Astronomical Society
• /
• v.46 no.2
• /
• pp.46.3-47
• /
• 2021

Recently a multilayer spectral inversion (MLSI) model has been proposed to infer the physical parameters of plasmas in the solar chromosphere. The inversion solves a three-layer radiative transfer model using the strong absorption line profiles, H alpha and Ca II 8542 Å, taken by the Fast Imaging Solar Spectrograph (FISS). The model successfully provides the physical plasma parameters, such as source functions, Doppler velocities, and Doppler widths in the layers of the photosphere to the chromosphere. However, it is quite expensive to apply the MLSI to a huge number of line profiles. For example, the calculating time is an hour to several hours depending on the size of the scan raster. We apply deep neural network (DNN) to the inversion code to reduce the cost of calculating the physical parameters. We train the models using pairs of absorption line profiles from FISS and their 13 physical parameters (source functions, Doppler velocities, Doppler widths in the chromosphere, and the pre-determined parameters for the photosphere) calculated from the spectral inversion code for 49 scan rasters (~2,000,000 dataset) including quiet and active regions. We use fully connected dense layers for training the model. In addition, we utilize a skip connection to avoid a problem of vanishing gradients. We evaluate the model by comparing the pairs of absorption line profiles and their inverted physical parameters from other quiet and active regions. Our result shows that the deep learning model successfully reproduces physical parameter maps of a scan raster observation per second within 15% of mean absolute percentage error and the mean squared error of 0.3 to 0.003 depending on the parameters. Taking this advantage of high performance of the deep learning model, we plan to provide the physical parameter maps from the FISS observations to understand the chromospheric plasma conditions in various solar features.

• Smart Structures and Systems
• /
• v.23 no.4
• /
• pp.359-371
• /
• 2019

As part of a structural health monitoring system, the relative geometric relationship between a ship and bridge has been recognized as important for bridge authorities and ship owners to avoid ship-bridge collision. This study proposes a novel computer vision method for the real-time geometric parameter identification of moving ships based on a single shot multibox detector (SSD) by using transfer learning techniques and monocular vision. The identification framework consists of ship detection (coarse scale) and geometric parameter calculation (fine scale) modules. For the ship detection, the SSD, which is a deep learning algorithm, was employed and fine-tuned by ship image samples downloaded from the Internet to obtain the rectangle regions of interest in the coarse scale. Subsequently, for the geometric parameter calculation, an accurate ship contour is created using morphological operations within the saturation channel in hue, saturation, and value color space. Furthermore, a local coordinate system was constructed using projective geometry transformation to calculate the geometric parameters of ships, such as width, length, height, localization, and velocity. The application of the proposed method to in situ video images, obtained from cameras set on the girder of the Wuhan Yangtze River Bridge above the shipping channel, confirmed the efficiency, accuracy, and effectiveness of the proposed method.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.745-755
• /
• 2023

The wire rope is an indispensable production machinery in coal mines. It is the main force-bearing equipment of the underground traction system. Accurate detection of wire rope defects and positions exerts an exceedingly crucial role in safe production. The existing defect detection solutions exhibit some deficiencies pertaining to the flexibility, accuracy and real-time performance of wire rope defect detection. To solve the aforementioned problems, this study utilizes the camera to sample the wire rope before the well entry, and proposes an object based on YOLOv5. The surface small-defect detection model realizes the accurate detection of small defects outside the wire rope. The transfer learning method is also introduced to enhance the model accuracy of small sample training. Herein, the enhanced YOLOv5 algorithm effectively enhances the accuracy of target detection and solves the defect detection problem of wire rope utilized in mine, and somewhat avoids accidents occasioned by wire rope damage. After a large number of experiments, it is revealed that in the task of wire rope defect detection, the average correctness rate and the average accuracy rate of the model are significantly enhanced with those before the modification, and that the detection speed can be maintained at a real-time level.

• International Journal of Computer Science & Network Security
• /
• v.23 no.9
• /
• pp.186-191
• /
• 2023

Neural Networks are widely used for huge variety of tasks solution. Machine Learning methods are used also for signal and time series analysis, including electrocardiograms. Contemporary wearable devices, both medical and non-medical type like smart watch, allow to gather the data in real time uninterruptedly. This allows us to transfer these data for analysis or make an analysis on the device, and thus provide preliminary diagnosis, or at least fix some serious deviations. Different methods are being used for this kind of analysis, ranging from medical-oriented using distinctive features of the signal to machine learning and deep learning approaches. Here we will demonstrate a neural network-based approach to this task by building an ensemble of 1D CNN classifiers and a final classifier of selection using logistic regression, random forest or support vector machine, and make the conclusions of the comparison with other approaches.

• Journal of Broadcast Engineering
• /
• v.27 no.2
• /
• pp.228-239
• /
• 2022

In this paper, we propose an unmanned vehicle scratch detection deep learning model for car sharing services. Conventional scratch detection models consist of two steps: 1) a deep learning module for scratch detection of images before and after rental, 2) a manual matching process for finding newly generated scratches. In order to build a fully automatic scratch detection model, we propose a one-step unmanned scratch detection deep learning model. The proposed model is implemented by applying transfer learning and fine-tuning to the deep learning model that detects changes in satellite images. In the proposed car sharing service, specular reflection greatly affects the scratch detection performance since the brightness of the gloss-treated automobile surface is anisotropic and a non-expert user takes a picture with a general camera. In order to reduce detection errors caused by specular reflected light, we propose a preprocessing process for removing specular reflection components. For data taken by mobile phone cameras, the proposed system can provide high matching performance subjectively and objectively. The scores for change detection metrics such as precision, recall, F1, and kappa are 67.90%, 74.56%, 71.08%, and 70.18%, respectively.