• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.04a
• /
• pp.647-650
• /
• 2002

The human faces do not have distinct features unlike other general objects. In general the features of eyes, nose and mouth which are first recognized when human being see the face are defined. These features have different characteristics depending on different human face. In this paper, We propose a face recognition algorithm using the hidden Markov model(HMM). In the preprocessing stage, we find edges of a face using the locally adaptive threshold scheme and extract features based on generic knowledge of a face, then construct a database with extracted features. In training stage, we generate HMM parameters for each person by using the forward-backward algorithm. In the recognition stage, we apply probability values calculated by the HMM to input data. Then the input face is recognized by the euclidean distance of face feature vector and the cross-correlation between the input image and the database image. Computer simulation shows that the proposed HMM algorithm gives higher recognition rate compared with conventional face recognition algorithms.

• Journal of Information Processing Systems
• /
• v.17 no.4
• /
• pp.754-771
• /
• 2021

In the task of continuous dimension emotion recognition, the parts that highlight the emotional expression are not the same in each mode, and the influences of different modes on the emotional state is also different. Therefore, this paper studies the fusion of the two most important modes in emotional recognition (voice and visual expression), and proposes a two-mode dual-modal emotion recognition method combined with the attention mechanism of the improved AlexNet network. After a simple preprocessing of the audio signal and the video signal, respectively, the first step is to use the prior knowledge to realize the extraction of audio characteristics. Then, facial expression features are extracted by the improved AlexNet network. Finally, the multimodal attention mechanism is used to fuse facial expression features and audio features, and the improved loss function is used to optimize the modal missing problem, so as to improve the robustness of the model and the performance of emotion recognition. The experimental results show that the concordance coefficient of the proposed model in the two dimensions of arousal and valence (concordance correlation coefficient) were 0.729 and 0.718, respectively, which are superior to several comparative algorithms.

• International Journal of Internet, Broadcasting and Communication
• /
• v.14 no.2
• /
• pp.55-62
• /
• 2022

A smart tourism chatbot is needed as a user interface to efficiently provide smart tourism services such as recommended travel products, tourist information, my travel itinerary, and tour guide service to tourists. We have been developed a smart tourism app and a smart tourism information system that provide smart tourism services to tourists. We also developed a smart tourism chatbot service consisting of khaiii morpheme analyzer, rule-based intention classification, and tourism information knowledge base using Neo4j graph database. In this paper, we develop the Korean and English smart tourism Name Entity (NE) datasets required for the development of the NER model using the pre-trained language models (PLMs) for the smart tourism chatbot system. We create the tourism information NER datasets by collecting source data through smart tourism app, visitJeju web of Jeju Tourism Organization (JTO), and web search, and preprocessing it using Korean and English tourism information Name Entity dictionaries. We perform training on the KoBERT-CRF NER model using the developed Korean and English tourism information NER datasets. The weight-averaged precision, recall, and f1 scores are 0.94, 0.92 and 0.94 on Korean and English tourism information NER datasets.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.477-480
• /
• 2022

반도체 성능 향상으로 신호를 전달하는 회로의 단위가 마이크로 미터에서 나노미터로 미세화되어 선폭(linewidth)이 점점 좁아지고 있다. 이러한 변화는 검출해야 할 불량의 크기가 작아지고, 정상 공정상태와 비정상 공정상태의 차이도 상대적으로 감소되어, 공정오차 및 공정조건의 허용범위가 축소되었음을 의미한다. 따라서 검출해야 할 이상징후 탐지가 더욱 어렵게 되어, 높은 정밀도와 해상도를 갖는 검사공정이 요구되고 있다. 이러한 이유로, 미세 공정변화를 파악할 수 있는 신규 검사 및 계측 공정이 추가되어 TAT(Turn-around Time)가 증가하게 되었고, 웨이퍼가 가공되어 완제품까지 도달하는데 필요한 공정시간이 증가하여 제조원가 상승의 원인으로 작용한다. 본 논문에서는 웨이퍼의 검계측 데이터가 아닌, 제조공정 과정에서 발생하는 다양한 센서 및 장비 데이터를 기반으로 웨이퍼 제조 결과가 양품인지 그렇지 않으면 불량인지 구별할 수 있는 가상계측 모델을 제안한다. 기계학습의 여러 알고리즘 중에서 다양한 장점을 갖는 XGBoost 알고리즘을 이용하여 예측모델을 구축하였고, 데이터 전처리(data-preprocessing), 주요변수 추출(feature selection), 모델 구축(model design), 모델 평가(model evaluation)의 순서로 연구를 수행하였다. 결과적으로 약 94% 이상의 정확성을 갖는 모형을 구축하는데 성공하였으나 더욱 높은 정확성을 확보하기 위해서는 반도체 공정과 관련된 Domain Knowledge 를 반영한 모델구축과 같은 추가적인 연구가 필요하다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.1
• /
• pp.169-177
• /
• 2020

In South Korea, the results of R&D in science and technology are submitted to the National Science and Technology Information Service (NTIS) in reports that have Korea national science and technology standard classification codes (K-NSCC). However, considering there are more than 2000 sub-categories, it is non-trivial to choose correct classification codes without a clear understanding of the K-NSCC. In addition, there are few cases of automatic document classification research based on the K-NSCC, and there are no training data in the public domain. To the best of our knowledge, this study is the first attempt to build a highly performing K-NSCC classification system based on NTIS report meta-information from the last five years (2013-2017). To this end, about 210 mid-level categories were selected, and we conducted preprocessing considering the characteristics of research report metadata. More specifically, we propose a convolutional neural network (CNN) technique using only task names and keywords, which are the most influential fields. The proposed model is compared with several machine learning methods (e.g., the linear support vector classifier, CNN, gated recurrent unit, etc.) that show good performance in text classification, and that have a performance advantage of 1% to 7% based on a top-three F1 score.

• Korean Journal of Cognitive Science
• /
• v.13 no.4
• /
• pp.1-8
• /
• 2002

As Internet grows, many web sites have been built, therefore much information has been registered. Because the web sites have more information, it is more difficult that the user can find the information wanted. Therefore, to get information that user wants easily, the full-text engine may be embedded to the web site. This paper is about developing the help conversational agent for a user to find the information that he wants through conversation with agent. The proposed method is based on the pattern matching of artificial intelligence, not natural language processing. If a user inputs any sentence, the help conversational agent responds to the sentence through preprocessing and pattern matching with knowledge. The knowledge is built with the XML format. With the approximate pattern matching, the agent picks up the appropriate response with some degree of similarities. At the experiment, some different sentences with the same meaning have been entered, then the agent recognized them as the same pattern, and it made a correct answer.

• The Korean Journal of Applied Statistics
• /
• v.37 no.1
• /
• pp.87-102
• /
• 2024

Digital transformation refers to the process by which a company or organization changes or innovates its existing business model or sales activities using digital technology. This requires the use of various digital technologies - cloud computing, IoT, artificial intelligence, etc. - to strengthen competitiveness in the market, improve customer experience, and discover new businesses. In addition, in order to derive knowledge and insight about the market, customers, and production environment, it is necessary to select the right data, preprocess the data to an analyzable state, and establish the right process for systematic analysis suitable for the purpose. The usefulness of such digital data is determined by the importance of pre-processing and the correct application of exploratory data analysis (EDA), which is useful for information and hypothesis exploration and visualization of knowledge and insights. In this paper, we reexamine the philosophy and basic concepts of EDA and discuss key visualization information, information expression methods based on the grammar of graphics, and the ACCENT principle, which is the final visualization review standard, for effective visualization.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.31 no.1
• /
• pp.74-82
• /
• 2008

The objective of this paper is to remove noises of image based on the heuristic noises filter and to extract a tumor region by using morphology techniques in breast ultrasound image. Similar objective studies have been conducted based on ultrasound image of high resolution. As a result, efficiency of noise removal is not fine enough for low resolution image. Moreover, when ultrasound image has multiple tumors, the extraction of ROI (Region Of Interest) is not accomplished or processed by a manual selection. In this paper, our method is done 4 kinds of process for noises removal and the extraction of ROI for solving problems of restrictive automated segmentation. First process is that pixel value is acquired as matrix type. Second process is a image preprocessing phase that is aimed to maximize a contrast of image and prevent a leak of personal information. In next process, the heuristic noise filter that is based on opinion of medical specialist is applied to remove noises. The last process is to extract a tumor region by using morphology techniques. As a result, the noise is effectively eliminated in all images and a extraction of tumor regions is possible though one ultrasound image has several tumors.

• Imaging Science in Dentistry
• /
• v.52 no.3
• /
• pp.239-244
• /
• 2022

Purpose: Despite the proliferation of numerous morphometric and anthropometric methods for sex identification based on linear, angular, and regional measurements of various parts of the body, these methods are subject to error due to the observer's knowledge and expertise. This study aimed to explore the possibility of automated sex determination using convolutional neural networks(CNNs) based on lateral cephalometric radiographs. Materials and Methods: Lateral cephalometric radiographs of 1,476 Iranian subjects (794 women and 682 men) from 18 to 49 years of age were included. Lateral cephalometric radiographs were considered as a network input and output layer including 2 classes(male and female). Eighty percent of the data was used as a training set and the rest as a test set. Hyperparameter tuning of each network was done after preprocessing and data augmentation steps. The predictive performance of different architectures (DenseNet, ResNet, and VGG) was evaluated based on their accuracy in test sets. Results: The CNN based on the DenseNet121 architecture, with an overall accuracy of 90%, had the best predictive power in sex determination. The prediction accuracy of this model was almost equal for men and women. Furthermore, with all architectures, the use of transfer learning improved predictive performance. Conclusion: The results confirmed that a CNN could predict a person's sex with high accuracy. This prediction was independent of human bias because feature extraction was done automatically. However, for more accurate sex determination on a wider scale, further studies with larger sample sizes are desirable.

• KIPS Transactions on Computer and Communication Systems
• /
• v.2 no.6
• /
• pp.279-290
• /
• 2013

In digital forensics log files have been stored as a form of large data for the purpose of tracing users' past behaviors. It is difficult for investigators to manually analysis the large log data without clues. In this paper, we propose a text mining technique for extracting intrusion logs from a large log set to recommend reliable evidences to investigators. In the training stage, the proposed method extracts intrusion association words from a training log set by using Apriori algorithm after preprocessing and the probability of intrusion for association words are computed by combining support and confidence. Robinson's method of computing confidences for filtering spam mails is applied to extracting intrusion logs in the proposed method. As the results, the association word knowledge base is constructed by including the weights of the probability of intrusion for association words to improve the accuracy. In the test stage, the probability of intrusion logs and the probability of normal logs in a test log set are computed by Fisher's inverse chi-square classification algorithm based on the association word knowledge base respectively and intrusion logs are extracted from combining the results. Then, the intrusion logs are recommended to investigators. The proposed method uses a training method of clearly analyzing the meaning of data from an unstructured large log data. As the results, it complements the problem of reduction in accuracy caused by data ambiguity. In addition, the proposed method recommends intrusion logs by using Fisher's inverse chi-square classification algorithm. So, it reduces the rate of false positive(FP) and decreases in laborious effort to extract evidences manually.