• IEMEK Journal of Embedded Systems and Applications
• /
• v.17 no.3
• /
• pp.129-138
• /
• 2022

Collecting and labeling sufficient training data, which is essential to deep learning-based visual inspection, is difficult for manufacturers to perform because it is very expensive. This paper presents a steel plate surface defect detection system with industrial-grade detection performance by training a small amount of steel plate surface images consisting of labeled and non-labeled data. To overcome the problem of lack of training data, we propose two data augmentation techniques: program-based augmentation, which generates defect images in a geometric way, and generative model-based augmentation, which learns the distribution of labeled data. We also propose a 4-step semi-supervised learning using pseudo labels and consistency training with fixed-size augmentation in order to utilize unlabeled data for training. The proposed technique obtained about 99% defect detection performance for four defect types by using 100 real images including labeled and unlabeled data.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.1
• /
• pp.69-74
• /
• 2005

In this paper, we suggested user assistant soft computing method for 3D effect optimization. In order to maximize 3D effect of image, intervals among cameras have to be set up properly according to distance between cameras and an object. Two data such as interval and distance was obtained to use in neural network as the data for learning. However, if the data for learning was obtained by only human's subjective views, it could be that the obtained data was not optimal for learning because the data had an accidental ewer To obtain optimal data lot learning, we added candidature data to obtained data through data analysis, and then selected the most proper data between the candidature data and the obtained data for learning in neural network. Usually, 3D effect of image was affected by both distance from an object to cameras and an object size. Therefore, we suggested fuzzy inference model which was able to represent two factors like distance and size. Candidature data was added by fuzzy model. In the simulation result, we verified that the mote the obtained data was affected by human's subjective views, the more effective the suggested system was.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.12
• /
• pp.4776-4794
• /
• 2020

We propose a rapid adaptive learning framework for streaming object detection, called EER-ASSL. The method combines the expected error reduction (EER) dependent rollback learning and the active semi-supervised learning (ASSL) for a rapid adaptive CNN detector. Most CNN object detectors are built on the assumption of static data distribution. However, images are often noisy and biased, and the data distribution is imbalanced in a real world environment. The proposed method consists of collaborative sampling and EER-ASSL. The EER-ASSL utilizes the active learning (AL) and rollback based semi-supervised learning (SSL). The AL allows us to select more informative and representative samples measuring uncertainty and diversity. The SSL divides the selected streaming image samples into the bins and each bin repeatedly transfers the discriminative knowledge of the EER and CNN models to the next bin until convergence and incorporation with the EER rollback learning algorithm is achieved. The EER models provide a rapid short-term myopic adaptation and the CNN models an incremental long-term performance improvement. EER-ASSL can overcome noisy and biased labels in varying data distribution. Extensive experiments shows that EER-ASSL obtained 70.9 mAP compared to state-of-the-art technology such as Faster RCNN, SSD300, and YOLOv2.

• The Journal of Korean Academic Society of Nursing Education
• /
• v.27 no.1
• /
• pp.80-90
• /
• 2021

Purpose: This study aimed to investigate nursing students' learning agility and confirm the mediating effect of self-leadership in the relationship between achievement motivation and learning agility. Methods: The study design was a descriptive survey design. The subjects were third- and fourth-year nursing students attending three universities in one region. Data were collected from November 28, 2019, to May 25, 2020, and a total of 202 data were collected using the scale of achievement motivation, self-leadership, and learning agility. Data analysis included frequency analysis, descriptive statistics, and Pearson's correlation coefficient using SPSS 25.0 statistics 25.0 software. The mediating effect of self-leadership was analyzed through regression analysis and bootstrapping using process macro ver. 3.4.1. Results: Self-leadership's partial mediating effect was confirmed in achievement motivation and learning agility. Achievement motivation was found to affect directly learning agility, with an indirect effect through self-leadership. Conclusion: The study results showed that nursing students could increase their learning agility through self-leadership improvement. Future research should focus on identifying the factors influencing nursing students' learning agility and develop and apply programs to improve learning agility.

• Journal of IKEEE
• /
• v.27 no.4
• /
• pp.537-547
• /
• 2023

In many domains, lack of data inhibits adoption of advanced machine learning models. Recently, Few-Shot Learning (FSL) has been actively studied to tackle this problem. Utilizing prior knowledge obtained through observations on related domains, FSL achieved significant performance with only a few samples. In this paper, we present a survey on FSL in terms of data augmentation, embedding and metric learning, and meta-learning. In addition to interesting researches, we also introduce major benchmark datasets. FSL is widely adopted in various domains, but we focus on image analysis in this paper.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.177-179
• /
• 2022

In the Google Deepmind Challenge match, Alphago defeated Korea's Sedol Lee (human) with 4 wins and 1 loss in the Go match. Finally, artificial intelligence is going beyond the use of human intelligence. The Korean government's budget for the Digital New Deal is 9 trillion won in 2022, and an additional 301 types of data construction projects for artificial intelligence learning will be secured. From 2023, the industrial paradigm will change with the use and application of learning of artificial intelligence in all fields of industry. This paper conducts research to utilize artificial intelligence algorithms. Focusing on the analysis and judgment of data in artificial intelligence learning, research on the appropriate target and scope of application of algorithms in artificial intelligence machine learning and deep learning learning is conducted. This study will provide basic data for artificial intelligence in the 4th industrial revolution technology and artificial intelligence robot use in the 5th industrial revolution technology.

• Journal of the Korean School Mathematics Society
• /
• v.5 no.2
• /
• pp.1-15
• /
• 2002

The purpose of this study came to start to be helpful in that the teachers in field produced evaluation subject about the mathematics and applied by analyzing the performance capacity for the students about the subject developed through the application of subject about the teaching-learning model that can utilize the learners' learning experience and the development of items that can evaluate it by utilizing various and concrete operational materials as an object of the first year at the middle schools. The concrete purpose of realizing it is as follows. 1. This study is to develop the teaching-learning model centered on activities adequate to utilize various and concrete operational materials. 2. This study is to develop evaluation data for utilizing the concrete operational materials by making construction through the application of developed teaching-learning model as possible. 3. This study is to present standards initiative for reliable and fair marking about the evaluation data and to analyze the students' practice capacity by using it effectively. For accomplishing the purpose of this research, this study is the work-oriented teaching and learning model using learning data that can be easily found around the surroundings and 8 evaluation programs in order that the experience-oriented learning based on circumstances learning among the learning models of constructivism. Also, it is to examine the result after applying that on a basis of 25 students, consisted of only on class in a year, in the first year at the small size of agricultural middle schools. The result of this study from this research is as follows. First, as a result of frequency survey, of response about the developed evaluation data it showed positive response more than 80% of all the items. The atmosphere of self-directed learning was produced because the developed evaluation data could induce activity-based learning naturally by stimulating the students' curiosity and promoting interest. Second, this study executed t-test for the result of questionnaire about the mathematics propensity, and there were significant differences in 19 items among 25 items. It presented that the application of data about the teaching-learning and evaluation directly using concrete operational materials to the constructional learning by level might have desirable affect on the students' mathematical propensity. It came to be a motive that could increase value recognition and interest on the subject of mathematics by investigating a mathematical principle rule and confirming it through activity learning using the concrete operational materials.

• Journal of Broadcast Engineering
• /
• v.25 no.2
• /
• pp.275-278
• /
• 2020

In this paper, we proposed a deep learning based character recognition system through virtual data generation. In order to secure the learning data that takes the largest weight in supervised learning, virtual data was created. Also, after creating virtual data, data generalization was performed to cope with various data by using augmentation parameter. Finally, the learning data composition generated data by assigning various values to augmentation parameter and font parameter. Test data for measuring the character recognition performance was constructed by cropping the text area from the actual image data. The test data was augmented considering the image distortion that may occur in real environment. Deep learning algorithm uses YOLO v3 which performs detection in real time. Inference result outputs the final detection result through post-processing.

• The Mathematical Education
• /
• v.61 no.2
• /
• pp.257-271
• /
• 2022

As information and communication technologies are being developed so rapidly, education research is actively conducted to provide optimal learning for each student using big data and artificial intelligence technology. In this study, using the mathematics learning data of elementary school 5th to 6th graders conducting blended mathematics classes, we tried to find out what factors predict mathematics academic achievement and developed an artificial intelligence model that predicts mathematics academic performance using the results. Math learning propensity, LMS data, and evaluation results of 205 elementary school students had analyzed with a random forest model. Confidence, anxiety, interest, self-management, and confidence in math learning strategy were included as mathematics learning disposition. The progress rate, number of learning times, and learning time of the e-learning site were collected as LMS data. For evaluation data, results of diagnostic test and unit test were used. As a result of the analysis it was found that the mathematics learning strategy was the most important factor in predicting low-achieving students among mathematics learning propensities. The LMS training data had a negligible effect on the prediction. This study suggests that an AI model can predict low-achieving students with learning data generated in a blended math class. In addition, it is expected that the results of the analysis will provide specific information for teachers to evaluate and give feedback to students.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.6
• /
• pp.301-307
• /
• 2023

Deep learning requires building a deep neural network, collecting a large amount of training data, and then training the built neural network for a long time. If training does not proceed properly or overfitting occurs, training will fail. When using deep learning tools that have been developed so far, it takes a lot of time to collect training data and learn. However, due to the rapid advent of the mobile environment and the increase in sensor data, the demand for real-time deep learning technology that can dramatically reduce the time required for neural network learning is rapidly increasing. In this study, a real-time deep learning system was implemented using an Arduino system equipped with a fine dust sensor. In the implemented system, fine dust data is measured every 30 seconds, and when up to 120 are accumulated, learning is performed using the previously accumulated data and the newly accumulated data as a dataset. The neural network for learning was composed of one input layer, one hidden layer, and one output. To evaluate the performance of the implemented system, learning time and root mean square error (RMSE) were measured. As a result of the experiment, the average learning error was 0.04053796, and the average learning time of one epoch was about 3,447 seconds.