• Phonetics and Speech Sciences
• /
• v.12 no.2
• /
• pp.29-37
• /
• 2020

Recently, the neural network-based deep learning algorithm has dramatically improved performance compared to the classical Gaussian mixture model based hidden Markov model (GMM-HMM) automatic speech recognition (ASR) system. In addition, researches on end-to-end (E2E) speech recognition systems integrating language modeling and decoding processes have been actively conducted to better utilize the advantages of deep learning techniques. In general, E2E ASR systems consist of multiple layers of encoder-decoder structure with attention. Therefore, E2E ASR systems require data with a large amount of speech-text paired data in order to achieve good performance. Obtaining speech-text paired data requires a lot of human labor and time, and is a high barrier to building E2E ASR system. Therefore, there are previous studies that improve the performance of E2E ASR system using relatively small amount of speech-text paired data, but most studies have been conducted by using only speech-only data or text-only data. In this study, we proposed a semi-supervised training method that enables E2E ASR system to perform well in corpus in different domains by using both speech or text only data. The proposed method works effectively by adapting to different domains, showing good performance in the target domain and not degrading much in the source domain.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.32 no.5
• /
• pp.987-995
• /
• 2022

Deep learning-based profiling side-channel analysis has been many proposed. Deep learning-based profiling analysis is a technique that trains the relationship between the side-channel information and the intermediate values to the neural network, then finds the secret key of the attack device using the trained neural network. Recently, cross-device profiling side channel analysis was proposed to consider the realistic deep learning-based profiling side channel analysis scenarios. However, it has a limitation in that attack performance is lowered if the profiling device and the attack device have not the same chips. In this paper, an environment in which the profiling device and the attack device have not the same chips is defined as the different-device, and a novel deep learning-based profiling side-channel analysis on different-device is proposed. Also, MCNN is used to well extract the characteristic of each data. We experimented with the six different boards to verify the attack performance of the proposed method; as a result, when the proposed method was used, the minimum number of attack traces was reduced by up to 25 times compared to without the proposed method.

• The Journal of Korean Academy of Orthopedic Manual Physical Therapy
• /
• v.12 no.1
• /
• pp.57-67
• /
• 2006

Musculoskeletal neck dysfunction syndromes are common in outpatient musculoskeletal pain practice. The underlying musculoskeletal and neurologic causes of pain are variable. In the management of these patients, it is important to accurately identify and treat these pain generators to optimize patient outcome. It is the purpose of this review to discuss three main categories of functional anatomy, the role of superficial/deep muscular system and the scientific evidence for optimal physical therapy intervention for cervical dysfunction. Specifically there is evidence of lowered microcirculation in the upper trapezius muscle, morphological signs of disturbed mitochondrial function which appears to be limited to type I fibers and an increased cross-sectional area of type I muscle fibers despite a lower capillary to fiber area ratio. In acute neck pain syndrome, changes in muscle activity of painful muscles may result from segmental and supraspinal inhibitory effects. Muscle activation is closely related to the control of joint movements and postures and it is difficult to separate the influence of the two components. Both the altered muscle recruitment patterns and altered kinematics appear to be a poor adaptation for pain of the head - neck region, as they are likely to result in increased compressive loading in the cervical spine, affecting muscles, articular structures such as zygapophyseal joints, connective tissues and neural tissues which are all peripheral generators of referred pain. The rectus capitus posterior minor muscle shows that it is one of the most important muscles of the suboccipital region. In this article, i reviewed the anatomy, neurophysiology, function and dysfunction as well as the treatment of cervical dysfunction.

• Journal of TMJ Balancing Medicine
• /
• v.12 no.1
• /
• pp.1-6
• /
• 2022

Objectives: For the deviation phenomenon occurring during the treatment process in temporo-mandibular balancing medicine (TBM), hypotheses were established regarding the cause and mechanism of formation from the perspective of neuro-science and material mechanics, and a verification method was proposed. Methods: The deviation phenomenon was theoretically analyzed based on the structure theories of material mechanics of the joint and the neurological pain mechanism. Results: Deviation occurs due to temporary yield by the accumulation of heterogeneous stress in the temporo-mandibular joint and the affected joint. Because the joint structures are corresponding with material mechanics showing compressive and tensile properties. The size of the deviation is expressed in terms of strain. The occlusal surface of the teeth is level with the axial joint. Since the magnitude of the deviation has a proportional relationship with the degree of abnormality of the temporo-mandibular joint, the magnitude of the deviation calculated by the balance measurement can be replaced by the strain. The major variables involved in the occurrence of deviations are the strength of joint structures and neurological conditions. Therefore plastic deformation and adaptation occur as a long-term depression of neural circuits is strengthened in different ways at different locations each time in various clinical situations. This is the reason why the sequence of the restoration process while correcting deviations is following reverse order of the accumulation in many layers in the muscular nervous system. Conclusions: From the above results, it can be inferred that the occurrence and correction of the deviations are corresponding with the plastic deformation and neuro-plasticity.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.1
• /
• pp.149-160
• /
• 1997

This paper is concerned with the problem of speaker-adaptive speech synthes is method using a mapped codebook designed by fuzzy mapping on FLVQ (Fuzzy Learning Vector Quantization). The FLVQ is used to design both input and reference speaker's codebook. This algorithm is incorporated fuzzy membership function into the LVQ(learning vector quantization) networks. Unlike the LVQ algorithm, this algorithm minimizes the network output errors which are the differences of clas s membership target and actual membership values, and results to minimize the distances between training patterns and competing neurons. Speaker Adaptation in speech synthesis is performed as follow;input speaker's codebook is mapped a reference speaker's codebook in fuzzy concepts. The Fuzzy VQ mapping replaces a codevector preserving its fuzzy membership function. The codevector correspondence histogram is obtained by accumulating the vector correspondence along the DTW optimal path. We use the Fuzzy VQ mapping to design a mapped codebook. The mapped codebook is defined as a linear combination of reference speaker's vectors using each fuzzy histogram as a weighting function with membership values. In adaptive-speech synthesis stage, input speech is fuzzy vector-quantized by the mapped codcbook, and then FCM arithmetic is used to synthesize speech adapted to input speaker. The speaker adaption experiments are carried out using speech of males in their thirties as input speaker's speech, and a female in her twenties as reference speaker's speech. Speeches used in experiments are sentences /anyoung hasim nika/ and /good morning/. As a results of experiments, we obtained a synthesized speech adapted to input speaker.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.40 no.4
• /
• pp.193-205
• /
• 2003

There has been so many research activities about robot soccer system in the many research fields, for example, intelligent control, communication, computer technology, sensor technology, image processing, mechatronics. Especially researchers research strategy for attacking in the field of strategy, and develop intelligent strategy. Then, soccer robots cannot defense completely and efficiently by using simple defense strategy. Therefore, intention extraction of attacker is needed for efficient defense. In this thesis, intention extractor of soccer robots is designed and developed based on FMMNN(Fuzzy Min-Max Neural networks ). First, intention for soccer robot system is defined, and intention extraction for soccer robot system is explained.. Next, FMMNN based intention extractor for soccer robot system is determined. FMMNN is one of the pattern classification method and have several advantages: on-line adaptation, short training time, soft decision. Therefore, FMMNN is suitable for soccer robot system having dynamic environment. Observer extracts attack intention of opponents by using this intention exactor, and this intention extractor is also used for analyzing strategy of opponent team. The capability of developed intention extractor is verified by simulation of 3 vs. 3 robot succor simulator. It was confirmed that the rates of intention extraction each experiment increase.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.21 no.3
• /
• pp.175-186
• /
• 2019

Efforts have been made to introduce the climate smart agriculture (CSA) for adaptation to future climate conditions, which would require collection and management of site specific meteorological data. The objectives of this study were to identify requirements for construction of agricultural meteorology information service system (AMISS) using technologies that lead to the fourth industrial revolution, e.g., internet of things (IoT), artificial intelligence, and cloud computing. The IoT sensors that require low cost and low operating current would be useful to organize wireless sensor network (WSN) for collection and analysis of weather measurement data, which would help assessment of productivity for an agricultural ecosystem. It would be recommended to extend the spatial extent of the WSN to a rural community, which would benefit a greater number of farms. It is preferred to create the big data for agricultural meteorology in order to produce and evaluate the site specific data in rural areas. The digital climate map can be improved using artificial intelligence such as deep neural networks. Furthermore, cloud computing and fog computing would help reduce costs and enhance the user experience of the AMISS. In addition, it would be advantageous to combine environmental data and farm management data, e.g., price data for the produce of interest. It would also be needed to develop a mobile application whose user interface could meet the needs of stakeholders. These fourth industrial revolution technologies would facilitate the development of the AMISS and wide application of the CSA.

• Journal of Intelligence and Information Systems
• /
• v.24 no.3
• /
• pp.1-19
• /
• 2018

Large amount of data is now available for research and business sectors to extract knowledge from it. This data can be in the form of unstructured data such as audio, text, and image data and can be analyzed by deep learning methodology. Deep learning is now widely used for various estimation, classification, and prediction problems. Especially, fashion business adopts deep learning techniques for apparel recognition, apparel search and retrieval engine, and automatic product recommendation. The core model of these applications is the image classification using Convolutional Neural Networks (CNN). CNN is made up of neurons which learn parameters such as weights while inputs come through and reach outputs. CNN has layer structure which is best suited for image classification as it is comprised of convolutional layer for generating feature maps, pooling layer for reducing the dimensionality of feature maps, and fully-connected layer for classifying the extracted features. However, most of the classification models have been trained using online product image, which is taken under controlled situation such as apparel image itself or professional model wearing apparel. This image may not be an effective way to train the classification model considering the situation when one might want to classify street fashion image or walking image, which is taken in uncontrolled situation and involves people's movement and unexpected pose. Therefore, we propose to train the model with runway apparel image dataset which captures mobility. This will allow the classification model to be trained with far more variable data and enhance the adaptation with diverse query image. To achieve both convergence and generalization of the model, we apply Transfer Learning on our training network. As Transfer Learning in CNN is composed of pre-training and fine-tuning stages, we divide the training step into two. First, we pre-train our architecture with large-scale dataset, ImageNet dataset, which consists of 1.2 million images with 1000 categories including animals, plants, activities, materials, instrumentations, scenes, and foods. We use GoogLeNet for our main architecture as it has achieved great accuracy with efficiency in ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Second, we fine-tune the network with our own runway image dataset. For the runway image dataset, we could not find any previously and publicly made dataset, so we collect the dataset from Google Image Search attaining 2426 images of 32 major fashion brands including Anna Molinari, Balenciaga, Balmain, Brioni, Burberry, Celine, Chanel, Chloe, Christian Dior, Cividini, Dolce and Gabbana, Emilio Pucci, Ermenegildo, Fendi, Giuliana Teso, Gucci, Issey Miyake, Kenzo, Leonard, Louis Vuitton, Marc Jacobs, Marni, Max Mara, Missoni, Moschino, Ralph Lauren, Roberto Cavalli, Sonia Rykiel, Stella McCartney, Valentino, Versace, and Yve Saint Laurent. We perform 10-folded experiments to consider the random generation of training data, and our proposed model has achieved accuracy of 67.2% on final test. Our research suggests several advantages over previous related studies as to our best knowledge, there haven't been any previous studies which trained the network for apparel image classification based on runway image dataset. We suggest the idea of training model with image capturing all the possible postures, which is denoted as mobility, by using our own runway apparel image dataset. Moreover, by applying Transfer Learning and using checkpoint and parameters provided by Tensorflow Slim, we could save time spent on training the classification model as taking 6 minutes per experiment to train the classifier. This model can be used in many business applications where the query image can be runway image, product image, or street fashion image. To be specific, runway query image can be used for mobile application service during fashion week to facilitate brand search, street style query image can be classified during fashion editorial task to classify and label the brand or style, and website query image can be processed by e-commerce multi-complex service providing item information or recommending similar item.

• Journal of Korean Theatre Studies Association
• /
• no.52
• /
• pp.319-357
• /
• 2014

Formation of 'body schema' is the start for actor to create role and becomes the root and the foundation of existing as a role on the stage. For this, an actor needs to form 'scheme of role' with escaping from own 'body schema.' 'Schema of role' is formed by acquiring through synthesizing daily basic actions, namely, walking, standing, sitting, hand stretching, bending, and touching. The body schema, which was made with simple and usual actions, has fundamental significance in a sense of becoming the body in which the past traces in a role are habituated while energy as a role flows. As for the process of forming body schema, an actor first needs to obtain the visualized materials like photo, magazine, picture and image available for seeing a role specifically and clearly based on what analyzed a character. An actor needs to have three-dimensional image available for always recalling it in the head during acting. To do this, image data available for fundamentally capturing routine actions along with body structure are still more useful. Next, the body schema is formed by interaction with environment. Thus, there is a need of passing through the two-time process of forming body schema. Firstly, the body schema is made on routine actions in a role as physical condition of a role in actor's own everyday life. Secondly, the body schema is made on routine actions available for moving efficiently and economically in line with the environment of performance. A theatrical stage is the temporal space of rhythm and rule different from routine space. What forms body schema immediately in the second phase without body schema in the first phase ultimately becomes what exists as actor's own body, not the body of a role. The body schema, which was formed as the second process, is what truly has identity as a role in the ontological aspect, comes to experience the oppositional force in muscle, a qualitative change in energy, and emotional agitation in the physical aspect, and experiences perception, thinking, volition, and even consciousness with the entire body in the cognitive dimension. Thus, the formation of body schema can be known to be just a method of changing even spiritual and emotional layer. Body schema cannot be made if there is no process of embodiment and habit. Embodiment and habit are not simply the repeated, empty and mechanical action in the body. But, habit itself has very important meanings for forming body schema for role creating. First, habit allows the body itself to learn and understand a meaning. Second, habit relies upon environment, thereby allowing an actor of making the habituated body schema to recognize environment. Third, habit makes the mind. The habituated body schema is just the mind and the ego of a person who possesses the body schema. Fourth, habit comes to experience the expansion in energy and the expansion in existence. It may be experienced through interrelation among actor's body, tool, and environment. Fifth, habit makes identity of the body. Hence, this just becomes what secures identity of a role. These implications of habit are the formation of body schema, which is maintained with the body of being remembered firmly through being closely connected with the process of neural adaptation. Finally, it sought for possibility of practice as one method of forming body schema for role creating through Deleuze's '-becoming' theory. As 'actual animal-becoming' is real '-becoming' of forming structural transformation in the physical dimension, it meets with what the formation of body schema pursues actuality and reality. This was explained with a concept as saying of 'all '-becoming' molecular' by Deleuze/Guattari. 'Animal of having imitated animal's characteristic- becoming' is formed by which the body schema relies upon environment. In this way, relationship among the body, tool and environment has influence even upon a change in consciousness, thinking, and emotion, thereby being able to be useful for forming body schema in a sense of possibly experiencing ultimately expansion in role, namely, expansion in existence.