• Korean Journal of Remote Sensing
• /
• v.37 no.1
• /
• pp.111-122
• /
• 2021

Building footprint extraction is an active topic in the domain of remote sensing, since buildings are a fundamental unit of urban areas. Deep convolutional neural networks successfully perform footprint extraction from optical satellite images. However, semantic segmentation produces coarse results in the output, such as blurred and rounded boundaries, which are caused by the use of convolutional layers with large receptive fields and pooling layers. The objective of this study is to generate visually enhanced building objects by directly extracting the vertices of individual buildings by combining instance segmentation and keypoint detection. The target keypoints in building extraction are defined as points of interest based on the local image gradient direction, that is, the vertices of a building polygon. The proposed framework follows a two-stage, top-down approach that is divided into object detection and keypoint estimation. Keypoints between instances are distinguished by merging the rough segmentation masks and the local features of regions of interest. A building polygon is created by grouping the predicted keypoints through a simple geometric method. Our model achieved an F1-score of 0.650 with an mIoU of 62.6 for building footprint extraction using the OpenCitesAI dataset. The results demonstrated that the proposed framework using keypoint estimation exhibited better segmentation performance when compared with Mask R-CNN in terms of both qualitative and quantitative results.

• Journal of IKEEE
• /
• v.25 no.4
• /
• pp.650-663
• /
• 2021

Subsurface topology estimation is an important factor in the geophysical survey. Electrical impedance tomography is one of the popular methods used for subsurface imaging. The EIT inverse problem is highly nonlinear and ill-posed; therefore, reconstructed conductivity distribution suffers from low spatial resolution. The subsurface region can be approximated as piece-wise separate regions with constant conductivity in each region; therefore, the conductivity estimation problem is transformed to estimate the shape and location of the layer boundary interface. Each layer interface boundary is treated as an open boundary that is described using front points. The subsurface domain contains multi-layers with very complex configurations, and, in such situations, conventional methods such as the modified Newton Raphson method fail to provide the desired solution. Therefore, in this work, we have implemented a 7-layer artificial neural network (ANN) as an inverse problem algorithm to estimate the front points that describe the multi-layer interface boundaries. An ANN model consisting of input, output, and five fully connected hidden layers are trained for interlayer boundary reconstruction using training data that consists of pairs of voltage measurements of the subsurface domain with three-layer configuration and the corresponding front points of interface boundaries. The results from the proposed ANN model are compared with the gravitational search algorithm (GSA) for interlayer boundary estimation, and the results show that ANN is successful in estimating the layer boundaries with good accuracy.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2022.11a
• /
• pp.57-58
• /
• 2022

Concrete mix design is used as essential data for the quality of concrete, analysis of structures, and stable use of sustainable structures. However, since most of the formulation design is established based on the experience of experts, there is a lack of data to base it on. are suffering Accordingly, in this study, the purpose of this study is to build a predictive model to use the concrete mixing factor as basic data for calculation using the DNN technique. As for the data set for DNN model learning, OPC and ternary concrete data were collected according to the presence or absence of admixture, respectively, and the model was separated for OPC and ternary concrete, and training was carried out. In addition, by varying the number of hidden layers of the DNN model, the prediction performance was evaluated according to the model structure. The higher the number of hidden layers in the model, the higher the predictive performance for the prediction of the mixing elements except for the compressive strength factor set as the output value, and the ternary concrete model showed higher performance than the OPC. This is expected because the data set used when training the model also affected the training.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.842-857
• /
• 2023

Currently, most sentiment classification models on microblogging platforms analyze sentence parts of speech and emoticons without comprehending users' emotional inclinations and grasping moral nuances. This study proposes a hybrid sentiment analysis model. Given the distinct nature of microblog comments, the model employs a combined stop-word list and word2vec for word vectorization. To mitigate local information loss, the TextCNN model, devoid of pooling layers, is employed for local feature extraction, while BiLSTM is utilized for contextual feature extraction in deep learning. Subsequently, microblog comment sentiments are categorized using a classification layer. Given the binary classification task at the output layer and the numerous hidden layers within BiLSTM, the Tanh activation function is adopted in this model. Experimental findings demonstrate that the enhanced TextCNN-BiLSTM model attains a precision of 94.75%. This represents a 1.21%, 1.25%, and 1.25% enhancement in precision, recall, and F1 values, respectively, in comparison to the individual deep learning models TextCNN. Furthermore, it outperforms BiLSTM by 0.78%, 0.9%, and 0.9% in precision, recall, and F1 values.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.1
• /
• pp.93-113
• /
• 2020

According to information theory, non-orthogonal transmission can achieve the multiple-user channel capacity with an onion-peeling like successive interference cancellation (SIC) based detection followed by a capacity approaching channel code. However, in multiple antenna system, due to the unideal characteristic of the SIC detector, the residual interference propagated to the next detection stage will significantly degrade the detection performance of spatial data layers. To overcome this problem, we proposed a modified power-domain non-orthogonal multiple access (P-NOMA) scheme joint designed with space-time coding for multiple input multiple output (MIMO) NOMA system. First, with proper power allocation for each user, inter-user signals can be separated from each other for NOMA detection. Second, a well-designed quasi-orthogonal space-time block code (QO-STBC) was employed to facilitate the SIC-based MIMO detection of spatial data layers within each user. Last, we proposed an optimization algorithm to assign channel coding rates to balance the bit error rate (BER) performance of those spatial data layers for each user. Link-level performance simulation results demonstrate that the proposed time-space-power domain joint transmission scheme performs better than the traditional P-NOMA scheme. Furthermore, the proposed algorithm is of low complexity and easy to implement.

• The KIPS Transactions:PartB
• /
• v.13B no.1 s.104
• /
• pp.15-20
• /
• 2006

This paper describes research on intelligent game character through performance enhancements of physics engine in computer games. The algorithm that recognizes the physics situation uses momentum back-propagation neural networks. Also, we present an experiment and its results, integration methods that display optimum performance based on the physics situation. In this experiment on integration methods, the Euler method was shown to produce the best results in terms of fps in a simulation environment with collision detection. Simulation with collision detection was shown similar fps for all three methods and the Runge-kutta method was shown the greatest accuracy. In the experiment on physics situation recognition, a physics situation recognition algorithm where the number of input layers (number of physical parameters) and output layers (destruction value for the master car) is fixed has shown the best performance when the number of hidden layers is 3 and the learning count number is 30,000. Since we tested with rigid bodies only, we are currently studying efficient physics situation recognition for soft body objects.

• The Korean Journal of Applied Statistics
• /
• v.32 no.5
• /
• pp.693-702
• /
• 2019

Deep learning has gained popularity for the classification and prediction task. Neural network layers become deeper as more data becomes available. Saturation is the phenomenon that the gradient of an activation function gets closer to 0 and can happen when the value of weight is too big. Increased importance has been placed on the issue of saturation which limits the ability of weight to learn. To resolve this problem, Glorot and Bengio (Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, 249-256, 2010) claimed that efficient neural network training is possible when data flows variously between layers. They argued that variance over the output of each layer and variance over input of each layer are equal. They proposed a method of initialization that the variance of the output of each layer and the variance of the input should be the same. In this paper, we propose a new method of establishing initialization by adopting truncated normal distribution and truncated cauchy distribution. We decide where to truncate the distribution while adapting the initialization method by Glorot and Bengio (2010). Variances are made over output and input equal that are then accomplished by setting variances equal to the variance of truncated distribution. It manipulates the distribution so that the initial values of weights would not grow so large and with values that simultaneously get close to zero. To compare the performance of our proposed method with existing methods, we conducted experiments on MNIST and CIFAR-10 data using DNN and CNN. Our proposed method outperformed existing methods in terms of accuracy.

• Journal of IKEEE
• /
• v.18 no.4
• /
• pp.593-602
• /
• 2014

One of the purposes of an artificial neural netowrk(ANNet) is to implement the largest number of functions as possible with the smallest number of nodes and layers. This paper presents a CoreNet which has a multi-leveled input value and a multi-leveled output value with a 2-layered ANNet, which is the basic structure of an ANNet. I have suggested an equation for calculating the capacity of the CoreNet, which has a p-leveled input and a q-leveled output, as $a_{p,q}={\frac{1}{2}}p(p-1)q^2-{\frac{1}{2}}(p-2)(3p-1)q+(p-1)(p-2)$. I've applied this CoreNet into the simulation model 1(5)-1(6), which has 5 levels of an input and 6 levels of an output with no hidden layers. The simulation result of this model gives, the maximum 219 convergences for the number of implementable functions using the cot(${\sqrt{x}}$) input leveling method. I have also shown that, the 27 functions are implementable by the calculation of weight values(w, ${\theta}$) with the multi-threshold lines in the weight space, which are diverged in the simulation results. Therefore the 246 functions are implementable in the 1(5)-1(6) model, and this coincides with the value from the above eqution $a_{5,6}(=246)$. I also show the implementable function numbering method in the weight space.

• Journal of Korean Institute of Industrial Engineers
• /
• v.40 no.6
• /
• pp.628-641
• /
• 2014

Stepwise efficiency improvement in data envelopment analysis (DEA)-based benchmarking is a realistic and effective method by which inefficient decision making units (DMUs) can choose benchmarks in a stepwise manner and, thereby, effect gradual performance improvement. Most of the previous research relevant to stepwise efficiency improvement has focused primarily on how to stratify DMUs into multiple layers and how to select immediate benchmark targets in leading levels for lagging-level DMUs. It can be said that the sequence of benchmark targets was constructed in a myopic way, which can limit its effectiveness. To address this issue, this paper proposes an optimization approach to the construction of a sequence of benchmarks in DEA-based benchmarking, wherein two optimization criteria are employed : similarity of input-output use patterns, and proximity of input-output use levels between DMUs. To illustrate the proposed method, we applied it to the benchmarking of 23 national universities in South Korea.

• Journal of Korean Academy of Nursing
• /
• v.26 no.2
• /
• pp.281-289
• /
• 1996

Neural networks have recently attracted considerable attention in the field of classification and other areas. The purpose of this study was to demonstrate an experiment using back-propagation neural network model applied to nursing diagnosis. The network's structure has three layers ; one input layer for representing signs and symptoms and one output layer for nursing diagnosis as well as one hidden layer. The first prototype of a nursing diagnosis system for patients with stomach cancer was developed with 254 nodes for the input layer and 20 nodes for the output layer of 20 nursing diagnoses, by utilizing learning data set collected from 118 patients with stomach cancer. It showed a hitting ratio of .93 when the model was developed with 20,000 times of learning, 6 nodes of hidden layer, 0.5 of momentum and 0.5 of learning coefficient. The system was primarily designed to be an aid in the clinical reasoning process. It was intended to simplify the use of nursing diagnoses for clinical practitioners. In order to validate the developed model, a set of test data from 20 patients with stomach cancer was applied to the diagnosis system. The data for 17 patients were concurrent with the result produced from the nursing diagnosis system which shows the hitting ratio of 85％. Future research is needed to develop a system with more nursing diagnoses and an evaluation process, and to expand the system to be applicable to other groups of patients.