• Smart Media Journal
• /
• v.1 no.2
• /
• pp.12-22
• /
• 2012

Today we are witnessing an increasingly widespread use of cameras in our lives for video surveillance, robot vision, and mobile phones. This has led to a renewed interest in computer vision in general and an on-going boom in human activity recognition in particular. Although not particularly fancy per se, human gait is inarguably the most common and frequent action. Early on this decade there has been a passing interest in human gait recognition, but it soon declined before we came up with a systematic analysis and understanding of walking motion. This paper presents a set of DBN-based models for the analysis of human gait in sequence of increasing complexity and modeling power. The discussion centers around HMM-based statistical methods capable of modeling the variability and incompleteness of input video signals. Finally a novel idea of extending the discrete state Markov chain with a continuous density function is proposed in order to better characterize the gait direction. The proposed modeling framework allows us to recognize pedestrian up to 91.67% and to elegantly decode out two independent gait components of direction and posture through a sequence of experiments.

• Phonetics and Speech Sciences
• /
• v.11 no.1
• /
• pp.41-49
• /
• 2019

An end-to-end speech recognition model consisting of a single integrated neural network model was recently proposed. The end-to-end model does not need several training steps, and its structure is easy to understand. However, it is difficult to understand how the model recognizes speech internally. In this paper, we visualized and analyzed the attention-based end-to-end model to elucidate its internal mechanisms. We compared the acoustic model of the BLSTM-HMM hybrid model with the encoder of the end-to-end model, and visualized them using t-SNE to examine the difference between neural network layers. As a result, we were able to delineate the difference between the acoustic model and the end-to-end model encoder. Additionally, we analyzed the decoder of the end-to-end model from a language model perspective. Finally, we found that improving end-to-end model decoder is necessary to yield higher performance.

• Journal of Digital Convergence
• /
• v.10 no.4
• /
• pp.341-346
• /
• 2012

Social network service is service that enables the human network to be built up on web. It is important to record users' information simply and establish the network with people based on the information to provide with the social network service effectively. But it is very troublesome work for the user to input his or her own information on the mobile environment. In this paper we suggested a system which classifies users' behavior using context and creates blogging sentences automatically after inferring the destination. For this, users' behavior is classified and the destination is inferred with the sequence matching method using Naive Bayes classification. Then sentences which are suitable for situation is created by arranging the processed context using the structure of 5W1H. The system was evaluated satisfaction degree by comparing the created sentences based on actually collected data with users' intension and got accuracy rate of 88.73%.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.46 no.3
• /
• pp.8-14
• /
• 2009

A Hidden Markov Model(HMM)-based information extraction method is proposed in this paper. The proposed extraction method is applied to extraction of products' prices. The input of the proposed IESHMM is the URLs of a search engine's interface, which contains the names of the product types. The output of the system is the list of extracted slots of each product: name, price, image, and URL. With the observation data set Maximum Likelihood algorithm and Baum-Welch algorithm are used for the training of HMM and The Viterbi algorithm is then applied to find the state sequence of the maximal probability that matches the observation block sequence. When applied to practical problems, the proposed HMM-based system shows improved results over a conventional method, PEWEB, in terms of recall ration and accuracy.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.2
• /
• pp.94-101
• /
• 2023

A deep neural network composed of encoders and decoders, such as U-Net, used for speech enhancement, concatenates the encoder to the decoder through skip-connection. Skip-connection helps reconstruct the enhanced spectrum and complement the lost information. The features of the encoder and the decoder connected by the skip-connection are incompatible with each other. In this paper, for complex-valued spectrum based speech enhancement, Self-Attention (SA) method is applied to skip-connection to transform the feature of encoder to be compatible with the features of decoder. SA is a technique in which when generating an output sequence in a sequence-to-sequence tasks the weighted average of input is used to put attention on subsets of input, showing that noise can be effectively eliminated by being applied in speech enhancement. The three models using encoder and decoder features to apply SA to skip-connection are studied. As experimental results using TIMIT database, the proposed methods show improvements in all evaluation metrics compared to the Deep Complex U-Net (DCUNET) with skip-connection only.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.8 no.2
• /
• pp.167-175
• /
• 1988

The principle objective of this study is to evaluate the fatigue behavior of structural steel components of highway bridges subjected to service stresses. The main aspects of this investigation are; 1) a measurement and statistical analysis of service stress cycles observed in highway bridge. 2) fatigue tests under equivalent constant-amplitude(CA) loading and simulated variable-amplitude(VA) loading 3) a evaluation of the fatigue behavior under VA-loading by eqivalent root mean cube (RMC) stress range. Theoretically, the RMC model is adequate in evaluation of fatigue behavior under VA-loading, because the regression coefficient (m) of crack growth rate is 3 approximately. The result of fatigue test shows that the RMC model is fitter than the current RMS model in fatigue evaluation under VA-loading. The interaction effects and sequence effects under VA-loading affect little fatigue life of structural components. As the transition rate of stress ranges is higher, the crack growth rate is higher.

• Korean Management Science Review
• /
• v.13 no.2
• /
• pp.25-48
• /
• 1996

Mixed Model Assembly Lines (MMALs) are increasingly used to produce differentiated products on a single assembly line without work-in-process storage, Usually, a typical MMAL consists of a number of (1) stations doing exactly the same operation on every job, (2) stations involving operations with different choices, and (3) stations offering operations that are not performed on every job, or that are performed on every job but with many options. For stations of the first type there is no sequencing problem at all. However, for the second type a set-up cost is incurred each time the operation switches from one choice to another. At the third type of stations, different models, requring different amounts and choices of assembly work, creates an uneven flow of work along the line and variations in the work load at these stations. When a subsequence of jobs requires more work load than the station can handle, it is necessary to help the operations at the station or to complete the work elsewhere. Therefore, a schedule which minimize the sum of set-up cost and utility work cost is desired. So this study has developed Fixed Random Ordering Rule (FROR), Fixed Ascending Ordering Rule (FAOR), Fixed Descending Ordering Rule, and Extended NHR (ENHR). ENHR is to choose optimal color ordering of each batch with NHR, and to decide job sequence of the batch with it, too. As the result of experiments, ENHR was the best heuristic algorithm. NHR is a new heuristic rule in which only the minimum addition of violations from both partial sequence and unassigned sequence at every branch could be considered. And this is a heuristic sequencing rule for the third type of stations at MMAL. This study developed one more heuristic algorithm to test the performance of NHR, which is named as Practical Heuristic Rule (PHR).

• Korean Journal of Microbiology
• /
• v.39 no.2
• /
• pp.114-117
• /
• 2003

In molecular biology, it is necessary to develop an easy and rapid method to identify a specific DNA sequence. Though Southern and Northern blot techniques have been used widely for the analysis of gene structure and function, those methods are inconvenient in the points that we need to control incubation temperature, time, and other parameters to get the final result. In this study, we report a new method for the rapid analysis of specific DNA sequence with the modification of an immunochromatographic method. The lateral flow DNA analysis strip is composed of a sample pad, a nitrocellulose membrane for the separation and propagation of analytes, and an absorption pad for the generation of capillary action. Capture DNA was immobilized on the membrane by UV cross-linking and target DNA was labeled with Cy-5 for signaling. The samples containing target DNA were applied onto the sample pad, incubated for 15 min for separation, and scanned with a GSI fluorescence scanner. Though the hybridization reaction occurs in a short time without any washing steps, there appears to be little cross hybridization between the different sequences. The result showed a possibility that the new method can be used for the rapid identification of specific DNA sequence among the samples.

• Journal of KIISE:Software and Applications
• /
• v.28 no.9
• /
• pp.634-648
• /
• 2001

Message Sequence Charts(MSC) has been used to describe the interactions of numerous concurrent tasks in telecommunication software. After the MSC specification is verified in requirement analysis phase, it can be used not only to synthesize state-based design models, but also to generate test sequences. Until now, the verification is accomplished by generating global state transition graph using the location information only. In this paper, we extend the condition statement of MSC to describe the activation condition of scenarios and the change of state variables, and propose an approach to construct global finite state machine (GFSM) using this information. The GFSM only includes feasible states and transitions of the system. We can generate the test sequences using the existing FSM-based test sequence generation technology.

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.2
• /
• pp.427-436
• /
• 2021

The purpose of this study is to implement a machine learning language model that learns literary texts. Literary texts have an important characteristic that pairs of question-and-answer are not frequently clearly distinguished. Also, literary texts consist of pronouns, figurative expressions, soliloquies, etc. They hinder the necessity of machine learning using literary texts by making it difficult to learn algorithms. Algorithms that learn literary texts can show more human-friendly interactions than algorithms that learn general sentences. For this goal, this paper proposes three text correction tasks that must be preceded in researches using literary texts for machine learning language model: pronoun processing, dialogue pair expansion, and data amplification. Learning data for artificial intelligence should have clear meanings to facilitate machine learning and to ensure high effectiveness. The introduction of special genres of texts such as literature into natural language processing research is expected not only to expand the learning area of machine learning, but to show a new language learning method.