• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.6
• /
• pp.1692-1705
• /
• 2024

In this paper, we present a method that integrates a Grammar Transducer as an external language model to enhance the accuracy of the pre-trained Korean End-to-end (E2E) Automatic Speech Recognition (ASR) model. The E2E ASR model utilizes the Connectionist Temporal Classification (CTC) loss function to derive hypothesis sentences from input audio. However, this method reveals a limitation inherent in the CTC approach, as it fails to capture language information from transcript data directly. To overcome this limitation, we propose a fusion approach that combines a clause-level n-gram language model, transformed into a Weighted Finite-State Transducer (WFST), with the E2E ASR model. This approach enhances the model's accuracy and allows for domain adaptation using just additional text data, avoiding the need for further intensive training of the extensive pre-trained ASR model. This is particularly advantageous for Korean, characterized as a low-resource language, which confronts a significant challenge due to limited resources of speech data and available ASR models. Initially, we validate the efficacy of training the n-gram model at the clause-level by contrasting its inference accuracy with that of the E2E ASR model when merged with language models trained on smaller lexical units. We then demonstrate that our approach achieves enhanced domain adaptation accuracy compared to Shallow Fusion, a previously devised method for merging an external language model with an E2E ASR model without necessitating additional training.

• International Journal of Computer Science & Network Security
• /
• v.23 no.12
• /
• pp.151-160
• /
• 2023

Breast cancer is the most dangerous and deadly form of cancer. Initial detection of breast cancer can significantly improve treatment effectiveness. The second most common cancer among Indian women in rural areas. Early detection of symptoms and signs is the most important technique to effectively treat breast cancer, as it enhances the odds of receiving an earlier, more specialist care. As a result, it has the possible to significantly improve survival odds by delaying or entirely eliminating cancer. Mammography is a high-resolution radiography technique that is an important factor in avoiding and diagnosing cancer at an early stage. Automatic segmentation of the breast part using Mammography pictures can help reduce the area available for cancer search while also saving time and effort compared to manual segmentation. Autoencoder-like convolutional and deconvolutional neural networks (CN-DCNN) were utilised in previous studies to automatically segment the breast area in Mammography pictures. We present Automatic SegmenAN, a unique end-to-end adversarial neural network for the job of medical image segmentation, in this paper. Because image segmentation necessitates extensive, pixel-level labelling, a standard GAN's discriminator's single scalar real/fake output may be inefficient in providing steady and appropriate gradient feedback to the networks. Instead of utilising a fully convolutional neural network as the segmentor, we suggested a new adversarial critic network with a multi-scale L1 loss function to force the critic and segmentor to learn both global and local attributes that collect long- and short-range spatial relations among pixels. We demonstrate that an Automatic SegmenAN perspective is more up to date and reliable for segmentation tasks than the state-of-the-art U-net segmentation technique.

• Horticultural Science & Technology
• /
• v.31 no.3
• /
• pp.289-298
• /
• 2013

Effects of curing treatments and storage temperature on the quality of Chinese yams (Dioscorea polystachya Turcz.) were investigated stepwise in three consecutive years for the optimization of postharvest handling procedures. Tuberous roots were harvested in early to mid November and cured under ambient or $29^{\circ}C$ heated air conditions for various periods according to the treatment conditions. Storage temperatures in the range of 0.5 to $7.5^{\circ}C$ were phased in to avoid chilling injury while examining storage potential from 4 to 7 months. As poststorage technology, short-term $60^{\circ}C$ hot-air exposure or low shelf temperature treatments were additively imposed. Curing treatments, especially heated air curing for 3-5 days tended to reduce the respiration and weight loss during storage while maintaining flesh firmness. Storage at $0.5^{\circ}C$ brought out typical chilling injury symptoms on the shelf with increases in respiration and lower flesh firmness by tissue breakdown resulting in the rapid loss of marketability. Optimum storage temperature appeared to be the $3-4^{\circ}C$ range which suppresses quality deterioration while avoiding chilling injury. Low shelf temperature seemed to be a necessary part of postharvest handling system to keep marketability through control of poststorage disorders such as rooting and decay. Overall results suggested that optimized postharvest program consisting of heated-air curing, storage at $3-4^{\circ}C$, and low shelf temperature could extend storage potential of Chinese yam to longer than 7 months.

• Journal of Nutrition and Health
• /
• v.31 no.9
• /
• pp.1446-1456
• /
• 1998

Bones are important parts in sustaining the shape of the body, but they are also metabolic organs which undergo bone remodeling by constant bone resorption and formation. Osteoporosis, the typical metabolic bone disease, is characterized by a reduction in bone mineral density (BMD). Women more than men are at risk fir osteoporosis-related fractures, especially in the lumbar spine, wrist, and hip region. Risk of fracture depends on one's BMD, which open determined by the peak bone mass value achieved at skeletal maturity and followed by subsequent age-and menopause-related bone loss. Genetic and environmental factors are known to play a key role in bone metabolism and diet is considered as one of the important environmental factors. The purpose of the present study was to assess the status of BMD and bone mineral content(BMC) to clarify the relationships between dietary intakes and the risk of osteoporosis in adult women in Taegu. Subjects were 130 healthy females in between 20 and 69 years of age. BMD and BMC of the lumbar spine(venebrae L2-4) of the subjects were measured by dual energy X-ray absorptiometry. The average age of the subjects was 47.4${\pm}$11.7 years old, the average weight was 57.2${\pm}$8.4kg, the average age of menarche was 16.6${\pm}$1.9 years old and the average age of menopause was 48.4${\pm}$5.3 years old. The nutrient intakes of the subjects measured by the convenient method were generally lower than the level of RDA. The result of nutrient intake assessed has shown that the average energy intake was 1701${\pm}$316kca1 which is 85.1% of the RDA and the average calcium intake was 485.4${\pm}$172.3mg which is 69.3% of the RDA. The intakes of protein, vitamin A, vitamin B$_1$, niacin were greater than the RDA, whereas the remaining nutrient intakes were lower than the RDA. The average BMD of the subjects was shown to be 1.06${\pm}$1.09g/$\textrm{cm}^2$. The highest BMD of 1.24${\pm}$0.14g/$\textrm{cm}^2$ was noticed in the subjects of 30s compared to 20s, 40s, 50s, 60s. The BMD values were compared by the relative body weight(R3W) of the menopause subjects, and it was found that the underweight group had significant lower BMB while the rest of the groups did not have any differences in BMD. The most strongly correlated nutrient with BMD among the menopause subjects appeared to be calcium. The women whose Ca intakes were higher than 500mg showed the significantly higher BMD than those with Ca intakes lower than 500mg. This study suggests that the most effective way to prevent osteoporosis and to reduce the incidences of fractures seems to be minimizing bone loss through the adequate intake of calcium as well as avoiding underweight, especially in menopausal women. (Korean J Nutrition 31(9) . 1446-1456, 1998)

• The Journal of Korean Society for School & Community Health Education
• /
• v.1 no.2
• /
• pp.105-125
• /
• 2000

The purpose of this study was to examine the influence of school safety accident on teacher's normal educational activities and to seek some desirable ways to cope with it. The subjects in this study were 351 class teachers randomly selected from Seoul and Kyonggi Province and surveyed from April through July, 2000. The conclusions were as follows; 1. Actual Condition of School Safety Accident 1) Approximately many teachers investigated had had an experience to suffer safety accident. Safety accident occurred most during break or class, but there was a significant difference according to service area. 2) Safety accident took place most in playground, and the most common cause was student's own carelessness, and the most widely occurred accident type was an injury. But there was no significant difference caused by the general characteristics of the teachers. 2. Influence of Safety Accident On Teacher's Educational Activities. 1) The largest reason they offered safety education was to ensure student safety. The greatest number of them had an opinion they would consider changing or giving up a planned normal educational activity if they recognized any possibilities of safety accident. There was a significant difference in this point according to gender and career. 2) They worried about possible safety accident most during field study, but there was a significant difference according to gender or presence or absence of safety accident experience. 3) The general characteristics of teacher produced a significant difference to an experience of avoiding educational activity due to psychological withdrawal, but safety accident experience didn't make any difference. 3. Minimization of Teacher Damage or Loss from Safety Accident. 1) The dominant opinion about teacher's small mistake for any occurrence of safety accident was that the responsibility should be escaped to maintain teacher's authority. For severe mistake, however, there were two different opinions at the same percentage: one was being exempted and the other was taking civil liability. 2) Establishing teacher insurance was preferred as a way to minimize teacher's economic loss from safety accident, but there was a significant difference according to gender. 3) The dominant opinion about the payment of insurance premium for safety accident was that it should be paid from school operating expenses.

• Geophysics and Geophysical Exploration
• /
• v.21 no.3
• /
• pp.183-197
• /
• 2018

In the delineation of geological structures using seismic exploration, it is very important to improve resolution of seismic data as well as accurate velocity model building and subsurface imaging. The resolution of seismic data can be enhanced by employing high-frequency energy sources or by applying deconvolution techniques in data processing. In marine seismic exploration, however, the main reason for degradation of resolution is the loss of specific frequency components due to ghosts. If we remove the ghosts, we can obtain broadband seismic data by avoiding frequency loss, and thus providing high-resolution subsurface images. Although ghosts can be properly filtered out in the data processing step, more effective broadband seismic technologies have been developing through the evolution of seismic instruments and the innovation of survey design. Overseas exploration companies developed brand-new configurations of receivers such as over/under streamer and variable-depth streamer, or ghost removal techniques using dual-sensor streamer to serve high-resolution imaging technologies. Unfortunately, neither broadband seismic instrument nor processing technique has been studied in Korea. In this paper, we introduce fundamental theories and current status of broadband seismic technologies to assist domestic researchers to study those technologies.

• Korean Journal of Veterinary Research
• /
• v.49 no.3
• /
• pp.257-263
• /
• 2009

Bovine mastitis is an important disease causing serious economic loss in dairy production and food poison in public health. Escherichia coli and Staphylococcus aureus are the major causative agents of bovine mastitis. These bacteria were found in milk and environmental condition such as feces, water, soil and so on. Bovine mastitis causative micro-organisms can survive in 1-2 weeks in feces and bed complexes. Low level of percentage of water content (PWC) of feces and bed complexes can reduce the spreading of bovine mastitis incidence from environmental contamination. In this study, we developed the fecal dehydrating system and determined the elimination rates of bovine mastitis causative agent from feces and bed complexes. To develop the fecal dehydrating system, the screw pressurized dehydrating system was used and the maximum rate of dehydrating was reached to 52% PWC using 90% PWC (wet base) of fecal and bed complexes. The elimination rates of the dehydrating system for E. coli and S. aureus were reached at 41.19 $\pm$ 7.84% to 62.55 $\pm$ 8.71% in various percentages of PWC of feces and bed complexes (80, 85 and 90%). These results suggested that the application of fecal dehydrating system would reduce the exposure of dairy cattle to bovine mastitis causing agents contaminated feces and bed complexes, and can be used for environmental bovine mastitis control avoiding misuse or abuse of chemical disinfectants and antibiotics in dairy farm.

• Korean Chemical Engineering Research
• /
• v.56 no.4
• /
• pp.555-560
• /
• 2018

The high temperature stability of a chloride mixture, $NaCl-KCl-ZnCl_2$ (NaKZn-Chloride), is investigated to evaluate its potential as a thermal storage material. A thermal storage media should maintain a stable thermal properties within the temperature range of heat storage. Results from an a priori experiment showed that the NaKZn-chloride is stable only up the much lower temperature, while its stability limit is reported to be $850^{\circ}C$ in the literature. This study aims to investigate if the thermal property is changed by the moisture absorbed in the heat storage material. The effect of moisture content on the thermal properties was measured. The results show that the melting point remains the same regardless of the amount of moisture absorbed. Meanwhile, the high temperature stability is lower for the moisture treated samples. The results of this work infer that the loss of a hygroscopic thermal storage media can be reduced by avoiding its contacts to moisture in designing high temperature thermal storage systems.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.10
• /
• pp.82-90
• /
• 2013

An Optical Network-on-Chip(ONoC) based on silicon photonics is one of promising technology for next generation exascale computing architectures. Recent active researches on ONoC focus on improving bandwidth further and avoiding path collisions by using wavelength division multiplexing (WDM). However, the number of wavelengths used for the WDM increases linearly as the number of Processing Element (PE) increases in existing ONoCs which adopt centralized routing architecture. The problem will also arises growing cost of optical devices such as light switches and light sources and limits the scalability of ONoC due to the sinal loss caused by interference of distinct light sources. In this paper, we proposes a distributed routing architecture for ONoC which is based on 2D-mesh structure using WDM technique and present a method that minimize the required number of wavelengths exploiting the connectivity of communication. In comparison with existing centralized routing architectures, results show reduction by 56% of the number of wavelengths and 21% of the number of optical switches in $8{\times}8$ networks.

• The KIPS Transactions:PartC
• /
• v.15C no.3
• /
• pp.149-156
• /
• 2008

JPEG2000 is the upcoming image coding standard that provides better compression rate and image quality compared with JPEG. Lazy-mode steganography guarantees the safe communication under the two information loss stages in JPEG2000. However, it causes the severe changes of the code-block noise variance sequence after embedding and that is detectable under the steganalysis using the Hilbert-Huang transform (HHT) based sequential analysis. In this paper, a JPEG2000 lazy-mode steganography method is presented. The code blocks which produce the sudden variation of the noise variance after embedding are estimated by calculating low precision code-block variance (LPV) and low precision code-block noise variance (LPNV). By avoiding those code-blocks from embedding, our algorithm preserves the sequence and makes stego images secure under the HHT-based steganalytic detection. In addition, it prevents a severe degradation of image quality by using JPEG2000 quality layer information. On various 2048 images, experiments are performed to show the effective reduction of the noise variation after message embedding and the stable performance against HHT-based steganalysis.