• Proceedings of the Korea Multimedia Society Conference
• /
• 1998.04a
• /
• pp.411-416
• /
• 1998

본 논문은 Wavelet 변환 영역에서 특징 벡터를 추출하여 ART2 신경회로망으로 실장 PCB 패턴을 인식하는 알고리즘을 제안한다. PCB 형태 정보는 Wavelet에 의해 주파수 영역으로 변환되고, 이들 계수 행렬로부터 특징 벡터로서 추출된다. ART2 신경회로망은 이러한 특징 벡터들을 입력벡터로 사용하여 인식한다. 실장 PCB 영상 55장을 사용하여 실험한 결고, 학습된 입력패턴은 물론 비학습 입력패턴에 대해서도 약 99%의 인식율을 얻었다. 또한 제안된 방법은 Wavelet 변환 영역사에서 수직, 수평, 대각선 정보만으로 특징 벡터를 구축함으로써 특징 추출 과정이 비교적 간단하고 특징 벡터의 수도 줄일 수 있어, 효과적인 특징벡터의 추출이 가능함을 보였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.2
• /
• pp.751-770
• /
• 2019

Action recognition is an essential task in computer vision due to the variety of prospective applications, such as security surveillance, machine learning, and human-computer interaction. The availability of more video data than ever before and the lofty performance of deep convolutional neural networks also make it essential for action recognition in video. Unfortunately, limited crafted video features and the scarcity of benchmark datasets make it challenging to address the multi-person action recognition task in video data. In this work, we propose a deep convolutional neural network-based Effective Hybrid Learning (EHL) framework for two-person interaction classification in video data. Our approach exploits a pre-trained network model (the VGG16 from the University of Oxford Visual Geometry Group) and extends the Faster R-CNN (region-based convolutional neural network a state-of-the-art detector for image classification). We broaden a semi-supervised learning method combined with an active learning method to improve overall performance. Numerous types of two-person interactions exist in the real world, which makes this a challenging task. In our experiment, we consider a limited number of actions, such as hugging, fighting, linking arms, talking, and kidnapping in two environment such simple and complex. We show that our trained model with an active semi-supervised learning architecture gradually improves the performance. In a simple environment using an Intelligent Technology Laboratory (ITLab) dataset from Inha University, performance increased to 95.6% accuracy, and in a complex environment, performance reached 81% accuracy. Our method reduces data-labeling time, compared to supervised learning methods, for the ITLab dataset. We also conduct extensive experiment on Human Action Recognition benchmarks such as UT-Interaction dataset, HMDB51 dataset and obtain better performance than state-of-the-art approaches.

• Korean Journal of Cognitive Science
• /
• v.18 no.2
• /
• pp.179-191
• /
• 2007

In a goal-oriented dialogue, spoken' intentions can be represented by domain actions that consist of pairs of a speech art and a concept sequence. The domain action prediction of user's utterance is useful to correct some errors that occur in a speech recognition process, and the domain action prediction of system's utterance is useful to generate flexible responses. In this paper, we propose a model to predict a domain action of the next utterance using a neural network. The proposed model predicts the next domain action by using a dialogue history vector and a current domain action as inputs of the neural network. In the experiment, the proposed model showed the precision of 80.02% in speech act prediction and the precision of 82.09% in concept sequence prediction.

• International Journal of Advanced Culture Technology
• /
• v.7 no.2
• /
• pp.195-200
• /
• 2019

Media art, which is a combination of media technology and art, is making a lot of progress in combination with AI, IoT and VR. This paper aims to meet people's needs by creating a video that simulates the dance moves of an object that users admire by using media art that features interactive interactions between users and works. The project proposed a universal image synthesis system that minimizes equipment constraints by utilizing a deep running-based Skeleton estimation system and one of the deep-running neural network structures, rather than a Kinect-based Skeleton image. The results of the experiment showed that the images implemented through the deep learning system were successful in generating the same results as the user did when they actually danced through inference and synthesis of motion that they did not actually behave.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.5
• /
• pp.1037-1045
• /
• 1996

In this paper, the parallel model for stereo camera is employed to find the heralding angle and the distance between a leading vehicle and the following vehicle, BART(Binocular Autonomous Research Team vehicle). Two TDNNs (Time Delay Neural Network) such as S-TDNN and A-TDNN are introduced to control BART. S-TDNN controls the speed of the following vehicle while A-TDNN controls the steering angle of BATR. A human drives BART to collect data which are used for training the said neural networks. The trained networks performed the vehicle tracking function satisfactorily under the same driving conditions performed by the human driver. The neural network approach has good portability which decreases costs and saves development time for the different types of vehicles.

• Smart Structures and Systems
• /
• v.22 no.2
• /
• pp.175-183
• /
• 2018

Time-series data often contain one of the most valuable pieces of information in many fields including manufacturing. Because time-series data are relatively cheap to acquire, they (e.g., vibration signals) have become a crucial part of big data even in manufacturing shop floors. Recently, deep-learning models have shown state-of-art performance for analyzing big data because of their sophisticated structures and considerable computational power. Traditional models for a machinery-monitoring system have highly relied on features selected by human experts. In addition, the representational power of such models fails as the data distribution becomes complicated. On the other hand, deep-learning models automatically select highly abstracted features during the optimization process, and their representational power is better than that of traditional neural network models. However, the applicability of deep-learning models to the field of prognostics and health management (PHM) has not been well investigated yet. This study integrates the "residual fitting" mechanism inherently embedded in the wavelet transform into the convolutional neural network deep-learning structure. As a result, the architecture combines a signal smoother and classification procedures into a single model. Validation results from rotor vibration data demonstrate that our model outperforms all other off-the-shelf feature-based models.

• Journal of Korea Multimedia Society
• /
• v.25 no.2
• /
• pp.440-449
• /
• 2022

In recent years, the air pollution and Air Quality Index (AQI) has been a pivotal point for researchers due to its effect on human health. Various research has been done in predicting the AQI but most of these studies, either lack dense temporal data or cover one or two air pollutant elements. In this paper, a hybrid Convolutional Neural approach integrated with recurrent neural network architecture (CNN-LSTM), is presented to find air pollution inference using a multivariate air pollutant elements dataset. The aim of this research is to design a robust and real-time air pollutant forecasting system by exploiting a neural network. The proposed approach is implemented on a 24-month dataset from Seoul, Republic of Korea. The predicted results are cross-validated with the real dataset and compared with the state-of-the-art techniques to evaluate its robustness and performance. The proposed model outperforms SVM, SVM-Polynomial, ANN, and RF models with 60.17%, 68.99%, 14.6%, and 6.29%, respectively. The model performs SVM and SVM-Polynomial in predicting O3 by 78.04% and 83.79%, respectively. Overall performance of the model is measured in terms of Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and the Root Mean Square Error (RMSE).

• International Journal of Computer Science & Network Security
• /
• v.22 no.3
• /
• pp.220-228
• /
• 2022

The directed tests produce an expectation model to assist the organization's heads and professionals with settling on the right and speedy choice. A directed deep learning strategy has been embraced and applied for SCADA information. In this paper, for the load shedding expectation overall power organization of Libya, a convolutional neural network with multi neurons is utilized. For contributions of the neural organization, eight convolutional layers are utilized. These boundaries are power age, temperature, stickiness and wind speed. The gathered information from the SCADA data set were pre-handled to be ready in a reasonable arrangement to be taken care of to the deep learning. A bunch of analyses has been directed on this information to get a forecast model. The created model was assessed as far as precision and decrease of misfortune. It tends to be presumed that the acquired outcomes are promising and empowering. For assessment of the outcomes four boundary, MSE, RMSE, MAPE and R2 are determined. The best R2 esteem is gotten for 1-overlap and it was 0.98.34 for train information and for test information is acquired 0.96. Additionally for train information the RMSE esteem in 1-overlap is superior to different Folds and this worth was 0.018.

• Journal of Korea Multimedia Society
• /
• v.22 no.6
• /
• pp.655-664
• /
• 2019

In this paper, we propose a novel way of combining multiple deep convolutional neural network (DCNN) architectures which work well for accurate video face identification by adopting a serial combination of 3D and 2D DCNNs. The proposed method first divides an input video sequence (to be recognized) into a number of sub-video sequences. The resulting sub-video sequences are used as input to the 3D DCNN so as to obtain the class-confidence scores for a given input video sequence by considering both temporal and spatial face feature characteristics of input video sequence. The class-confidence scores obtained from corresponding sub-video sequences is combined by forming our proposed class-confidence matrix. The resulting class-confidence matrix is then used as an input for learning 2D DCNN learning which is serially linked to 3D DCNN. Finally, fine-tuned, serially combined DCNN framework is applied for recognizing the identity present in a given test video sequence. To verify the effectiveness of our proposed method, extensive and comparative experiments have been conducted to evaluate our method on COX face databases with their standard face identification protocols. Experimental results showed that our method can achieve better or comparable identification rate compared to other state-of-the-art video FR methods.

• Journal of Sensor Science and Technology
• /
• v.29 no.1
• /
• pp.19-26
• /
• 2020

Electrocardiogram (ECG) classification has become an essential task of modern day wearable devices, and can be used to detect cardiovascular diseases. State-of-the-art Artificial Intelligence (AI)-based ECG classifiers have been designed using various artificial neural networks (ANNs). Despite their high accuracy, ANNs require significant computational resources and power. Herein, three different ANNs have been compared: multilayer perceptron (MLP), convolutional neural network (CNN), and spiking neural network (SNN) only for the ECG classification. The ANN model has been developed in Python and Theano, trained on a central processing unit (CPU) platform, and deployed on a PYNQ-Z2 FPGA board to validate the model using a Jupyter notebook. Meanwhile, the hardware accelerator is designed with Overlay, which is a hardware library on PYNQ. For classification, the MIT-BIH dataset obtained from the Physionet library is used. The resulting ANN system can accurately classify four ECG types: normal, atrial premature contraction, left bundle branch block, and premature ventricular contraction. The performance of the ECG classifier models is evaluated based on accuracy and power. Among the three AI algorithms, the SNN requires the lowest power consumption of 0.226 W on-chip, followed by MLP (1.677 W), and CNN (2.266 W). However, the highest accuracy is achieved by the CNN (95%), followed by MLP (76%) and SNN (90%).