• 전자공학회논문지B
• /
• 제28B권5호
• /
• pp.382-393
• /
• 1991

In this paper, we propose a method for recognizing printed Korean characters using the Preceding Layer Driven multi-layer perceptron. The new learning algorithm which assigns the weight values to an integer and makes use of the transfer function as the step function was presented to design the hardware. We obtained 522 Korean character-image as an experimental object through scanner with 600DPI resolution. The preprocessing for feature extraction of Korean character is the separation of individual character, noise elimination smoothing, thinnig, edge point extraction, branch point extraction, and stroke segmentation. The used feature data are the number of edge points and their shapes, the number of branch points, and the number of strokes with 8 directions.

• 대한임베디드공학회논문지
• /
• 제16권5호
• /
• pp.171-177
• /
• 2021

It is difficult to accurately extract features of pedestrian because the pedestrian is photographed at a crosswalk using a camera positioned higher than the pedestrian. In addition, it is more difficult to extract features when a part of the pedestrian's body is covered by an umbrella or parasol or when the pedestrian is holding an object. Representative methods to solve this problem include Object Detection, Instance Segmentation, and Pose Estimation. Among them, this study intends to use the Pose Estimation method. In particular, we intend to increase the recognition rate of pedestrians in crosswalks by maintaining the image resolution through HigherHRNet and applying the foot estimation technique. Finally, we show the superiority of the proposed method by applying and analyzing several data sets covered by body parts to the existing method and the proposed method.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2022년도 추계학술발표대회
• /
• pp.818-820
• /
• 2022

The keyword 'healing' is essential to the competitive society and culture of Koreans. In addition, as the time at home increases due to COVID-19, the demand for indoor healing services has increased. Therefore, this thesis analyzes the user's facial expression so that people can receive various 'customized' healing services indoors, and based on this, provides lighting, ASMR, video recommendation service, and facial expression recording service.The user's expression was analyzed by applying the ensemble algorithm to the expression prediction results of various CNN models after extracting only the face through object detection from the image taken by the user.

• 한국빅데이터학회지
• /
• 제4권2호
• /
• pp.1-12
• /
• 2019

인공지능 기술이 발전하면서 이미지, 음성, 텍스트 등 다양한 분야에 적용되고 있으며, 데이터가 충분한 경우 기존 기법들에 비해 좋은 결과를 보인다. 주식시장은 경제, 정치와 같은 많은 변수에 의해 영향을 받기 때문에, 주식 가격의 움직임 예측은 어려운 과제로 알려져 있다. 다양한 기계학습 기법과 인공지능 기법을 이용하여 주가 패턴을 연구하여 주가의 등락을 예측하려는 시도가 있어왔다. 본 연구는 딥러닝 기법 중 컨볼루셔널 뉴럴 네트워크(CNN)를 기반으로 주가 패턴 예측률 향상을 위한 데이터 증강 방안을 제안한다. CNN은 컨볼루셔널 계층을 통해 이미지에서 특징을 추출하여 뉴럴 네트워크를 이용하여 이미지를 분류한다. 따라서, 본 연구는 주식 데이터를 캔들스틱 차트 이미지로 만들어 CNN을 통해 패턴을 예측하고 분류하고자 한다. 딥러닝은 다량의 데이터가 필요하기에, 주식 차트 이미지에 다양한 데이터 증강(Data Augmentation) 방안을 적용하여 분류 정확도를 향상 시키는 방법을 제안한다. 데이터 증강 방안으로는 차트를 랜덤하게 변경하는 방안과 차트에 가우시안 노이즈를 적용하여 추가 데이터를 생성하였으며, 추가 생성된 데이터를 활용하여 학습하고 테스트 집합에 대한 분류 정확도를 비교하였다. 랜덤하게 차트를 변경하여 데이터를 증강시킨 경우의 분류 정확도는 79.92%였고, 가우시안 노이즈를 적용하여 생성된 데이터를 가지고 학습한 경우의 분류 정확도는 80.98%이었다. 주가의 다음날 상승/하락으로 분류하는 경우에는 60분 단위 캔들 차트가 82.60%의 정확도를 기록하였다.

• 한국가구학회지
• /
• 제27권4호
• /
• pp.362-374
• /
• 2016

The ability of 'creative thinking' is one of the most important fact for designers to complete their design process. It is not only the overall concept of the process but also all details related to each steps of the process such as planning product strategies, research, finding ways to solve problems image development, broadening ideas, etc. It is, however, not easy to creative as designers by natural learning through experiences unless they have geniuses for creative designing. Many experienced designers in the field are even in dilemmas being not be able to differentiate from copying other designs. This study is focused on the researches on the works of famous designers, architects and artists who have shown great creativities throw their works or processes. It is aimed to define their ways of thinking which could be fundamental motives for the creativeness, various behavioral methods to solve problems in creative ways, technical approaches to improve creativities, and overall intuition in the design working processes.

• Journal of Information Processing Systems
• /
• 제17권2호
• /
• pp.399-410
• /
• 2021

Aiming at the problem that complex external variables in natural scenes have a greater impact on facial expression recognition results, a facial expression recognition method based on two-stream convolutional neural network is proposed. The model introduces exponentially enhanced shared input weights before each level of convolution input, and uses soft attention mechanism modules on the space-time features of the combination of static and dynamic streams. This enables the network to autonomously find areas that are more relevant to the expression category and pay more attention to these areas. Through these means, the information of irrelevant interference areas is suppressed. In order to solve the problem of poor local robustness caused by lighting and expression changes, this paper also performs lighting preprocessing with the lighting preprocessing chain algorithm to eliminate most of the lighting effects. Experimental results on AFEW6.0 and Multi-PIE datasets show that the recognition rates of this method are 95.05% and 61.40%, respectively, which are better than other comparison methods.

• 전기학회논문지
• /
• 제68권1호
• /
• pp.153-158
• /
• 2019

In this paper, we analyze the infrared feature for the small coast targets according to the surrounding environment for autonomous flight device equipped with an infrared imaging sensor and we propose Cross Duality of Neural Network (CR-DuNN) method which can classify the target and clutter in coastal environment. In coastal environment, there are various property according to diverse change of air temperature, sea temperature, deferent seasons. And small coast target have various infrared feature according to diverse change of environment. In this various environment, it is very important thing that we analyze and classify targets from the clutters to improve target detection accuracy. Thus, we propose infrared feature learning algorithm through LSTM neural network and also propose CR-DuNN algorithm that integrate LSTM prediction network with Du-CNN classification network to classify targets from the clutters.

• 한국수학교육학회지시리즈D:수학교육연구
• /
• 제17권2호
• /
• pp.133-152
• /
• 2013

In this research, we investigated students' understanding of the definitions of sequence of continuous functions and its uniform convergence. We selected three female and three male students out of the senior class of a university and conducted questionnaire surveys 4 times. We examined students' concept images of sequence of continuous functions and its uniform convergence and also how they approach to the right concept definitions for those through several progressive questions. Furthermore, we presented some suggestions for effective teaching-learning for the sequences of continuous functions.

• 한국가시화정보학회지
• /
• 제19권1호
• /
• pp.68-73
• /
• 2021

In this study, the detection of rubber o-ring defects was carried out using k-fold cross validation and Support Vector Machine (SVM) algorithm. The data process was carried out in 3 steps. First, we proceeded with a frame alignment to eliminate unnecessary regions in the learning and secondly, we applied gray-scale changes for computational reduction. Finally, data processing was carried out using image augmentation to prevent data overfitting. After processing data, SVM algorithm was used to obtain normal and defect detection accuracy. In addition, we applied the SVM algorithm through the k-fold cross validation method to compare the classification accuracy. As a result, we obtain results that show better performance by applying the k-fold cross validation method.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 춘계학술대회
• /
• pp.78-79
• /
• 2022

본 논문에서는 RGB-D 카메라를 이용하여 획득한 다수 사용자의 정규화된 스켈레톤의 신체 비율 분석을 통해 각 사용자의 구분 및 위치를 추적하는 기법을 제안한다. 이를 위해 3D 포인트 클라우드로부터 각 사용자의 3D 스켈레톤을 추출한 뒤 신체 비율 정보를 저장한다. 이후 저장된 신체 비율 정보를 전체 프레임에서 출력된 신체 비율 데이터와 유사도를 비교하여 전체 영상에서의 사용자 구분 및 위치추적 알고리즘을 제안한다.