• KIISE Transactions on Computing Practices
• /
• v.23 no.7
• /
• pp.417-423
• /
• 2017

Recently, many embedded devices that have the computing capability required for deep learning have become available; hence, many new applications using these devices are emerging. However, these embedded devices have an architecture different from that of PCs and high-performance servers. In this paper, we propose a method that improves the performance of deep-learning framework by considering the architecture of an embedded device that shares memory between the CPU and the GPU. The proposed method is implemented in Caffe, an open-source deep-learning framework, and is evaluated on an NVIDIA Jetson TK1 embedded device. In the experiment, we investigate the image recognition performance of several state-of-the-art deep-learning networks, including AlexNet, VGGNet, and GoogLeNet. Our results show that the proposed method can achieve significant performance gain. For instance, in AlexNet, we could reduce image recognition latency by about 33% and energy consumption by about 50%.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.22 no.6
• /
• pp.577-583
• /
• 2011

Radar performance test is one of the major steps for radar system design. However, it is restricted by time and cost when radar performance tests are performed with opportunity targets. So various simulated target generators are developed and used to evaluate radar performance. To simulate the target's range, most of simulated target generators are developed with optical line or DRFM(Digital RF Memory) technique but there are many restrictions such as limit of range imitation and test scenario because of their original usage. In this paper, DDL(Digital Delay Line) module for development of simulated target generator is designed with precise range simulation and easily embodiment feature. And pulse recognition and delay gap detection technique are used to simulate the time delay without distortions. Developed DDL module performances are verified through their performance tests and test results are described in this paper.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.4
• /
• pp.717-724
• /
• 2020

Due to the important role played by emotion in human interaction, affective computing is dedicated in trying to understand and regulate emotion through human-aware artificial intelligence. By understanding, emotion mental diseases such as depression, autism, attention deficit hyperactivity disorder, and game addiction will be better managed as they are all associated with emotion. Various studies for emotion recognition have been conducted to solve these problems. In applying machine learning for the emotion recognition, the efforts to reduce the complexity of the algorithm and improve the accuracy are required. In this paper, we investigate emotion Electroencephalogram (EEG) feature reduction and classification using Stack AutoEncoder (SAE) and Long-Short-Term-Memory/Recurrent Neural Networks (LSTM/RNN) classification respectively. The proposed method reduced the complexity of the model and significantly enhance the performance of the classifiers.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.7
• /
• pp.1759-1772
• /
• 2023

Chinese named entity recognition (NER) is a challenging work that seeks to find, recognize and classify various types of information elements in unstructured text. Due to the Chinese text has no natural boundary like the spaces in the English text, Chinese named entity identification is much more difficult. At present, most deep learning based NER models are developed using a bidirectional long short-term memory network (BiLSTM), yet the performance still has some space to improve. To further improve their performance in Chinese NER tasks, we propose a new NER model, IDCNN-BiLSTM-Highway, which is a combination of the BiLSTM, the iterated dilated convolutional neural network (IDCNN) and the highway network. In our model, IDCNN is used to achieve multiscale context aggregation from a long sequence of words. Highway network is used to effectively connect different layers of networks, allowing information to pass through network layers smoothly without attenuation. Finally, the global optimum tag result is obtained by introducing conditional random field (CRF). The experimental results show that compared with other popular deep learning-based NER models, our model shows superior performance on two Chinese NER data sets: Resume and Yidu-S4k, The F1-scores are 94.98 and 77.59, respectively.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.39 no.1
• /
• pp.97-107
• /
• 2002

In this paper, we propose the realization of on-line handwritten Chinese character recognition for mobile personal digital assistants (PDA). We focus on the development of an algorithm having a high recognition performance under the restriction that PDA requires small memory storage and less computational complexity in comparison with PC. Therefore, we use index matching method having computational advantage for fast recognition and we suggest a unit reconstruction method to minimize the memory size to store the character models and to accomodate the various changes in stroke order and stroke number of each person in handwriting Chinese characters. We set up standard model consisting of 1800 characters using a set of pre-defined units. Input data are measured by similarity among candidate characters selected on the basis of stroke numbers and region features after preprocessing and feature extracting. We consider 1800 Chinese characters adopted in the middle and high school in Korea. We take character sets of five person, written in printed style, irrespective of stroke ordering and stroke numbers. As experimental results, we obtained an average recognition time of 0.16 second per character and the successful recognition rate of 94.3% with MIPS R4000 CPU in PDA.

• The Journal of the Acoustical Society of Korea
• /
• v.11 no.3
• /
• pp.40-52
• /
• 1992

In this paper, we propose the speaker adaptive word recognition method using a mapped membership function, in order to absorb a fluctuation owing to personal difference which is a problem of speaker independent speech recognition. In the training procedure of this method, the mapped membership function is made with the fuzzy theory introducded into a mapped codebook, between an unknown speaker's spectrum pattern and a standard speaker's one. In the recognition procedure, an input pattern of an unknown speaker is reconstructed to the pattern which is adapted to that of a standard speaker by the mapped membership function. To show the validity of this method, word recognition experiments are carried out using 28 DDD area names. The recognition rate of the conventional speaker-adaptive method using a mapped codebook by VQ is 64.9[%], and that made by a fuzzy VQ is 76.2[%]. Throughout the experiment using a mapped membership function, we can achieve 95.4[%] recognition rate. This shows that our proposed method is more excellent in recognition performance. Moreover, this method doesn't need an iterative training procedure to make the mapped membership function, and memory capacity and computation requirements for this method are reduced to 1/30 and 1/500 time of those for the conventional method using a mapped codebook, respectively.

• Anxiety and mood
• /
• v.10 no.1
• /
• pp.37-43
• /
• 2014

Objective : Social Phobia is associated with extensive disability and reduced quality of life. The concept of 'social self' is a representation of the self-reflected in the eyes of others, and is recruited during self-face recognition, which is closely related to self-esteem. The aim of this study was to identify the relationship of face recognition for self and others using measures of social anxiety and self-esteem in patients with social phobia. Methods : Twenty-seven patients with social phobia and twenty-three normal controls were evaluated with scales of self-esteem, depression, anxiety and other psychiatric symptoms. All participants completed the self-face recognition task. Nine self-faces, nine other faces and eighty-one morphed faces were presented randomly for each trial. The participants were instructed to make a decision as to whether the stimuli were self-face or not. The responses and reaction times were recorded during the task. Results : There were no group differences of the morphing composition at the recognition start point as self-face. In patients with social phobia, the mean reaction time at the start point of recognizing as a self-face was 1,037.6 ms, which was significantly longer than that of normal controls (911.3 ms, p<0.05). Patients with social phobia showed a significant negative correlation between the mean reaction time and the severity of depression when the stimuli were recognized as a self-face (r=-0.421, p<0.05). Conclusion : A difficulty in attention rather than avoidance may be an important factor of face recognition in patients with social phobia. When considering self-face recognition in such patients, many factors, such as anxiety, depression, working memory and theory of mind, need to be considered.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.4
• /
• pp.883-898
• /
• 1996

In this paper, we propose improved method to recognize the shape for enhancing the quality of the pattern recognition system by compressing the source images. In the proposed method, we reduced the data amount by skeletonizing the source images using mathematical morphology, and then matched patterns after accomplishing the translation and scale normalization, and rotation invariance on the transformed images. Through the scale normalization, it was possible for the shape recognition at minimum amount of the pixel by giving the weight to the skeleton pixel. As the source images was replaced by the skeleton images, it was possible to reduce the amount of data and computational loads dramatically, and so become much faster even with a smaller memory capacity. Through the experiment, we investigated the optimum scale factor and good result was proved when realizing the pattern recognition system.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.5C
• /
• pp.627-631
• /
• 2004

An efficient algorithm for clustering of GPDFs(Gaussian Probability Density Functions) in a speech recognition model is proposed in this paper. The proposed algorithm is based on CNN with the divergence as its distance measure and is applied to a speech recognition. The algorithm is compared with conventional Dk-means(Divergence-based k-means) algorithm in CDHMM(Continuous Density Hidden Markov Model). The results show that it can reduce about 31.3％ of GPDFs over Dk-means algorithm without suffering any recognition performance. When compared with the case that no clustering is employed and full GPDFs are used, the proposed algorithm can save about 61.8％ of GPDFs while preserving the recognition performance.

• The Research Journal of the Costume Culture
• /
• v.15 no.6
• /
• pp.996-1007
• /
• 2007

This study was focused on recognitive reaction. The purposes of this study was to analyze the fashion brands through correlation analysis between top of mind, recall, recognition, impact index and consumer behavior, and to identify the graveyard brand, niche brand and core brand. 33 questions about 20 fashion brands were asked to 442 males and females from the middle school students to age of 40. Data were analyzed mean, standard deviation, frequency, and correlation by using SPSS 12.0. The results were as follows: 1. Top of mind, recall and recognition affected recognizing the brands and including evoked setting, but it didn't lead the customer to purchase the brand. 2. Although top of the mind and recall are high, the percentage of purchasing the brand is relatively low if a consumer doesn't own the brand. 3. Brands 'B', 'L', 'PF', 'D' and 'BM' were represented as niche Brands which had high recognition and memory. 4. Brands 'TB', 'I', 'EN', 'ML', 'E' could be Graveyard brands that need special management. 5. Brands with the high impact index were 'A', 'T', 'I', 'C' and 'B'. These brands were recognized as the core brands by consumers.