• Smart Media Journal
• /
• v.11 no.11
• /
• pp.17-24
• /
• 2022

With the spread of smart speakers based on voice recognition technology and deep learning technology, not only non-disabled people, but also the blind or physically handicapped can easily control home appliances such as lights and TVs through voice by linking home network services. This has greatly improved the quality of life. However, in the case of speech-impaired people, it is impossible to use the useful services of the smart speaker because they have inaccurate pronunciation due to articulation or speech disorders. In this paper, we propose a personalized voice classification technique for the speech-impaired to use for some of the functions provided by the smart speaker. The goal of this paper is to increase the recognition rate and accuracy of sentences spoken by speech-impaired people even with a small amount of data and a short learning time so that the service provided by the smart speaker can be actually used. In this paper, data augmentation and one cycle learning rate optimization technique were applied while fine-tuning ResNet18 model. Through an experiment, after recording 10 times for each 30 smart speaker commands, and learning within 3 minutes, the speech classification recognition rate was about 95.2%.

• Journal of Korea Water Resources Association
• /
• v.55 no.spc1
• /
• pp.1211-1222
• /
• 2022

Abnormal condition inevitably occurs during operation of water distribution system (WDS) and requires the isolation of certain areas using isolation valves. In general, the determination of the optimal location of isolation valves considered minimization of hydraulic failures as isolation of certain areas causes a change in hydraulic states (e.g., flow direction, velocity, pressure, etc.). Water quality failure can also be induced by changes in hydraulics, which have not been considered for isolation valve system design. Therefore, this study proposes a new isolation valve system design methodology to prevent unexpected water quality failure events. The new methodology considers flow direction change ratio (FDCR), which accounts for flow direction changes after isolation of the area, as a constraint while reliability is used as the objective function. The optimal design model has been applied to a synthetic grid network and the results are compared with the traditional design approach. Results show that considering FDCR can eliminate flow direction changes while average pressure and coefficient of variation of pressure, velocity, and hydraulic geodesic index (HGI) outperform compared to the traditional design approach. The proposed methodology is expected to be a useful approach to minimizing unexpected consequences by traditional design approaches.

• Journal of Broadcast Engineering
• /
• v.27 no.1
• /
• pp.3-12
• /
• 2022

Image deblurring aims to remove image blur, which can be generated while shooting the pictures by the movement of objects, camera shake, blurring of focus, and so forth. With the rise in popularity of smartphones, it is common to carry portable digital cameras daily, so image deblurring techniques have become more significant recently. Originally, image deblurring techniques have been studied using traditional optimization techniques. Then with the recent attention on deep learning, deblurring methods based on convolutional neural networks have been actively proposed. However, most of them have been developed while focusing on better performance. Therefore, it is not easy to use in real situations due to the speed of their algorithms. To tackle this problem, we propose a novel deep learning-based deblurring algorithm that can be operated in real-time on HD resolution. In addition, we improved the training and inference process and could increase the performance of our model without any significant effect on the speed and the speed without any significant effect on the performance. As a result, our algorithm achieves real-time performance by processing 33.74 frames per second at 1280×720 resolution. Furthermore, it shows excellent performance compared to its speed with a PSNR of 29.78 and SSIM of 0.9287 with the GoPro dataset.

• Journal of the Korea Computer Graphics Society
• /
• v.29 no.3
• /
• pp.85-92
• /
• 2023

Recently, interest in artificial intelligence has increased due to the development of artificial intelligence fields such as ChatGPT and self-driving cars. However, there are still many unknown elements in training process of artificial intelligence, so that optimizing the model requires more time and effort than it needs. Therefore, there is a need for a tool or methodology that can analyze the weight changes during the training process of artificial intelligence and help out understatnding those changes. In this research, I propose a visualization system which helps people to understand the accumulated weight changes. The system calculates the weights for each training period to accumulates weight changes and stores accumulated weight changes to plot them in 3D space. This research will allow us to explore different aspect of artificial intelligence learning process, such as understanding how the model get trained and providing us an indicator on which hyperparameters should be changed for better performance. These attempts are expected to explore better in artificial intelligence learning process that is still considered as unknown and contribute to the development and application of artificial intelligence models.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.171-173
• /
• 2022

The 4th industrial revolution technology is developing people's lives more efficiently. GIS provided on the Internet services such as traffic information and time information makes people getting more quickly to destination. National geographic information service(NGIS) and each local government are making basic data to investigate SOC accessibility for analyzing optimal point. To construct the shortest distance, the accessibility from the starting point to the arrival point is analyzed. Applying road network map, the starting point and the ending point, the shortest distance, the optimal accessibility is calculated by using Dijkstra algorithm. The analysis information from multiple starting points to multiple destinations was required more than 3 steps of manual analysis to decide the position for the optimal point, within about 0.1% error. It took more time to process the many-to-many (M×N) calculation, requiring at least 32G memory specification of the computer. If an optimal proximity analysis service is provided at a desired location more versatile, it is possible to efficiently analyze locations that are vulnerable to business start-up and living facilities access, and facility selection for the public.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.4
• /
• pp.687-692
• /
• 2024

The energy crisis is emerging as a serious problem around the world. In the case of Korea, there is great interest in energy efficiency research related to industrial complexes, which use more than 53% of total energy and account for more than 45% of greenhouse gas emissions in Korea. One of the studies is a study on saving energy through sharing facilities between factories using the same utility in an industrial complex called a virtual energy network plant and through transactions between energy producing and demand factories. In such energy-saving research, data collection is very important because there are various uses for data, such as analysis and prediction. However, existing systems had several shortcomings in reliably collecting time series data. In this study, we propose an intelligent IIoT platform to improve it. The intelligent IIoT platform includes a preprocessing system to identify abnormal data and process it in a timely manner, classifies abnormal and missing data, and presents interpolation techniques to maintain stable time series data. Additionally, time series data collection is streamlined through database optimization. This paper contributes to increasing data usability in the industrial environment through stable data collection and rapid problem response, and contributes to reducing the burden of data collection and optimizing monitoring load by introducing a variety of chatbot notification systems.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.9
• /
• pp.1-7
• /
• 2024

In this paper, we proposes a Convolutional Neural Networks(CNN) equipped with Batch Normalization(BN) for handwritten digit recognition training the MNIST dataset. Aiming to surpass the performance of LeNet-5 by LeCun et al., a 6-layer neural network was designed. The proposed model processes 28×28 pixel images through convolution, Max Pooling, and Fully connected layers, with the batch normalization to improve learning stability and performance. The experiment utilized 60,000 training images and 10,000 test images, applying the Momentum optimization algorithm. The model configuration used 30 filters with a 5×5 filter size, padding 0, stride 1, and ReLU as activation function. The training process was set with a mini-batch size of 100, 20 epochs in total, and a learning rate of 0.1. As a result, the proposed model achieved a test accuracy of 99.22%, surpassing LeNet-5's 99.05%, and recorded an F1-score of 0.9919, demonstrating the model's performance. Moreover, the 6-layer model proposed in this paper emphasizes model efficiency with a simpler structure compared to LeCun et al.'s LeNet-5 (7-layer model) and the model proposed by Ji, Chun and Kim (10-layer model). The results of this study show potential for application in real industrial applications such as AI vision inspection systems. It is expected to be effectively applied in smart factories, particularly in determining the defective status of parts.

• International conference on construction engineering and project management
• /
• 2024.07a
• /
• pp.1307-1307
• /
• 2024

Recent research on heavy equipment has been conducted for the purposes of enhanced safety, productivity improvement, and carbon neutrality at construction sites. A sensor-based approach is being explored to monitor the location and movements of heavy equipment in real time. However, it poses significant challenges in terms of time and cost as multiple sensors should be installed on numerous heavy equipment at construction sites. In addition, there is a limitation in identifying the collaboration or interference between two or more heavy equipment. In light of this, a vision-based deep learning approach is being actively conducted to effectively respond to various working conditions and dynamic environments. To enhance the performance of a vision-based activity recognition model, it is essential to secure a sufficient amount of training datasets (i.e., video datasets collected from actual construction sites). However, due to safety and security issues at construction sites, there are limitations in adequately collecting training dataset under various situations and environmental conditions. In addition, the videos feature a sequence of multiple activities of heavy equipment, making it challenging to clearly distinguish the boundaries between preceding and subsequent activities. To address these challenges, this study proposed a domain adaptation in vision-based transfer learning for automated excavator activity recognition utilizing 3D ResNet (residual deep neural network). Particularly, this study aimed to identify the optimal depth of 3D ResNet (i.e., the number of layers of the feature extractor) suitable for domain adaptation via fine-tuning process. To achieve this, this study sought to evaluate the activity recognition performance of five 3D ResNet models with 18, 34, 50, 101, and 152 layers, which used two consecutive videos with multiple activities (5 mins, 33 secs and 10 mins, 6 secs) collected from actual construction sites. First, pretrained weights from large-scale datasets (i.e., Kinetic-700 and Moment in Time (MiT)) in other domains (e.g., humans, animals, natural phenomena) were utilized. Second, five 3D ResNet models were fine-tuned using a customized dataset (14,185 clips, 60,606 secs). As an evaluation index for activity recognition model, the F1 score showed 0.881, 0.689, 0.74, 0.684, and 0.569 for the five 3D ResNet models, with the 18-layer model performing the best. This result indicated that the activity recognition models with fewer layers could be advantageous in deriving the optimal weights for the target domain (i.e., excavator activities) when fine-tuning with a limited dataset. Consequently, this study identified the optimal depth of 3D ResNet that can maintain a reliable performance in dynamic and complex construction sites, even with a limited dataset. The proposed approach is expected to contribute to the development of decision-support systems capable of systematically managing enhanced safety, productivity improvement, and carbon neutrality in the construction industry.

• Journal of Korean Society of Transportation
• /
• v.22 no.3 s.74
• /
• pp.145-157
• /
• 2004

The cycle length design model of the Korean traffic responsive signal control systems is devised to vary a cycle length as a response to changes in traffic demand in real time by utilizing parameters specified by a system operator and such field information as degrees of saturation of through phases. Since no explicit guideline is provided to a system operator, the system tends to include ambiguity in terms of the system optimization. In addition, the cycle lengths produced by the existing model have yet been verified if they are comparable to the ones minimizing delay. This paper presents the studies conducted (1) to find shortcomings embedded in the existing model by comparing the cycle lengths produced by the model against the ones minimizing delay and (2) to propose a new direction to design a cycle length minimizing delay and excluding such operator oriented parameters. It was found from the study that the cycle lengths from the existing model fail to minimize delay and promote intersection operational conditions to be unsatisfied when traffic volume is low, due to the feature of the changed target operational volume-to-capacity ratio embedded in the model. The 64 different neural network based cycle length design models were developed based on simulation data surrogating field data. The CORSIM optimal cycle lengths minimizing delay were found through the COST software developed for the study. COST searches for the CORSIM optimal cycle length minimizing delay with a heuristic searching method, a hybrid genetic algorithm. Among 64 models, the best one producing cycle lengths close enough to the optimal was selected through statistical tests. It was found from the verification test that the best model designs a cycle length as similar pattern to the ones minimizing delay. The cycle lengths from the proposed model are comparable to the ones from TRANSYT-7F.

• Journal of Bio-Environment Control
• /
• v.28 no.2
• /
• pp.95-103
• /
• 2019

Temperature and relative humidity are important factors in crop cultivation and should be properly controlled for improving crop yield and quality. In order to control the environment accurately, we need to predict how the environment will change in the future. The objective of this study was to predict air temperature and relative humidity at a future time by using a multilayer perceptron (MLP). The data required to train MLP was collected every 10 min from Oct. 1, 2016 to Feb. 28, 2018 in an eight-span greenhouse ($1,032m^2$) cultivating mango (Mangifera indica cv. Irwin). The inputs for the MLP were greenhouse inside and outside environment data, and set-up and operating values of environment control devices. By using these data, the MLP was trained to predict the air temperature and relative humidity at a future time of 10 to 120 min. Considering typical four seasons in Korea, three-day data of the each season were compared as test data. The MLP was optimized with four hidden layers and 128 nodes for air temperature ($R^2=0.988$) and with four hidden layers and 64 nodes for relative humidity ($R^2=0.990$). Due to the characteristics of MLP, the accuracy decreased as the prediction time became longer. However, air temperature and relative humidity were properly predicted regardless of the environmental changes varied from season to season. For specific data such as spray irrigation, however, the numbers of trained data were too small, resulting in poor predictive accuracy. In this study, air temperature and relative humidity were appropriately predicted through optimization of MLP, but were limited to the experimental greenhouse. Therefore, it is necessary to collect more data from greenhouses at various places and modify the structure of neural network for generalization.