• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.3
• /
• pp.115-122
• /
• 2019

For Korean phoneme recognition, Hidden Markov-Gaussian Mixture model(HMM-GMM) or hybrid models which combine artificial neural network with HMM have been mainly used. However, current approach has limitations in that such models require force-aligned corpus training data that is manually annotated by experts. Recently, researchers used neural network based phoneme recognition model which combines recurrent neural network(RNN)-based structure with connectionist temporal classification(CTC) algorithm to overcome the problem of obtaining manually annotated training data. Yet, in terms of implementation, these RNN-based models have another difficulty in that the amount of data gets larger as the structure gets more sophisticated. This problem of large data size is particularly problematic in the Korean language, which lacks refined corpora. In this study, we introduce CTC algorithm that does not require force-alignment to create a Korean phoneme recognition model. Specifically, the phoneme recognition model is based on convolutional neural network(CNN) which requires relatively small amount of data and can be trained faster when compared to RNN based models. We present the results from two different experiments and a resulting best performing phoneme recognition model which distinguishes 49 Korean phonemes. The best performing phoneme recognition model combines CNN with 3hop Bidirectional LSTM with the final Phoneme Error Rate(PER) at 3.26. The PER is a considerable improvement compared to existing Korean phoneme recognition models that report PER ranging from 10 to 12.

• Journal of Animal Science and Technology
• /
• v.64 no.1
• /
• pp.1-9
• /
• 2022

Equilume light masks had no impact on hastening the resumption of estrous cyclicity in mares maintained in outdoor pastures on the mainland of Korea due to the cold weather conditions. Jeju Island is a major horse-breeding site in Korea and is warmer than the mainland during the winter season. Therefore, the primary objective of this study was to explore the efficiency of the Equilume light mask on the resumption of seasonal estrous cycles in Thoroughbred mares on Jeju Island. A total of 20 nonpregnant mares were randomly divided into the Equilume light mask (n = 9) and stable lighting (n =11) groups. The experiment was performed at seven different horse-breeding farms located on Jeju Island from November 15, 2020, to February 15, 2021. The mares were exposed to the respective lights from 16:00 to 23:00. Follicle size and uterine edema were measured by ultrasound scanning. Body condition scores (BCS) were also monitored during the experiment. Statistical analysis was conducted using the SAS and SPSS software, and p-values of < 0.05 were considered statistically significant. Two of the nine (22.2%) mares in the Equilume light mask group and three of the 11 (27.28%) mares in the stable lighting group were still cycling in December and January, which were considered as all-year-round cycling mares. On February 15, there was no difference between groups in the resumption of early seasonal estrus cycle, which was determined by follicles > 25 mm in addition to uterine edema. All mares in the Equilume light mask group and five of the eight mares (62.5%) in the stable lighting group had resumed cycling. Interestingly, six of the seven mares (87.5%) in the Equilume light mask and four of eight mares (50%) in the stable lighting group had already ovulated on February 15 (p > 0.05), as determined by the presence of a recent corpus luteum. No difference was observed in BCS and uterine edema between groups (p > 0.05). In conclusion, the Equilume light mask can be an effective approach to induce early seasonal estrus cycles of mares in Jeju Island, and it also enhances the efficiency of farm management by reducing labor.

• Phonetics and Speech Sciences
• /
• v.15 no.3
• /
• pp.75-82
• /
• 2023

Automatic speech recognition (ASR) has been revolutionized with deep learning-based approaches, among which self-supervised learning methods have proven to be particularly effective. In this study, we aim to enhance the performance of OpenAI's Whisper model, a multilingual ASR system on the Korean language. Whisper was pretrained on a large corpus (around 680,000 hours) of web speech data and has demonstrated strong recognition performance for major languages. However, it faces challenges in recognizing languages such as Korean, which is not major language while training. We address this issue by fine-tuning the Whisper model with an additional dataset comprising about 1,000 hours of Korean speech. We also compare its performance against a Transformer model that was trained from scratch using the same dataset. Our results indicate that fine-tuning the Whisper model significantly improved its Korean speech recognition capabilities in terms of character error rate (CER). Specifically, the performance improved with increasing model size. However, the Whisper model's performance on English deteriorated post fine-tuning, emphasizing the need for further research to develop robust multilingual models. Our study demonstrates the potential of utilizing a fine-tuned Whisper model for Korean ASR applications. Future work will focus on multilingual recognition and optimization for real-time inference.

• Clinical and Experimental Reproductive Medicine
• /
• v.12 no.2
• /
• pp.25-37
• /
• 1985

Four points on the basal body temperature (B.B.T.) curve was correlated with the estimated time of ovulation, as determined by serial ultrasound in 50 induced menstrual cycles from 22 subjects. The time of ovulation was estimated by measuring the maximal diameter of follicles and observing the morphologic changes within the ovary from follicle to corpus luteum. The results were as following; 1. The diameter of the follicle measured at the day before disappearance was 21.1 mm on an average (S.D.: 2.14). The average follicular growth for 4 days before ovulation was measured at a rate of 2.8 mm/day, and rapid growth of follicle was observed 3.1 mm/day at the day before. 2. The changes associated with rupture of the follicles were the followings, in order of frequency; decrease in size(94%), disappearance of follicles(64%), fluid in the Cul-de-Sac(26%) and increased internal echoes(16%). 3. Only 20 of 50 cycles, exhibited a BBT dip and correlated with the estimated time of ovulation by ultrasound in 2 of which cases(10%). BBT nadir, 30 of 50 cycles, correlated in 5(16.7%). The first day of hyperthermic plateau(FDHP) and BBT coverline was exhibited in all cycles, correlated in 41(82%) and 35(70%) cases. 4. The relationship between the diameter of dominant dominant follicle, measured by ultrasound, and the basal body temperature curve were as following. During cycles in which dip was observed on the BBT curve, the follicular diameter were 10.5${\pm}$2.12 mm on 4 days prior to the point (D-4), and 12.5${\pm}$2.12 mm (D-3), 15.5${\pm$2.12 mm (D-2), 17.0${\pm}$1.41 mm (D-1) and 21.5${\pm}$2.12 mm just prior to the dip (D-0). In the nadir; 9.6${\pm}$1.67 mm (N-4), 12.8${\pm}$1.79 mm (N-3), 16.2${\pm}$1.92 mm (N-2), 18.2${\pm}$2.17 mm (N-1) and 21.4${\pm}$2.61 mm (N-0). In the First day of Hyperthemic Plateau (FDHP); 9.8${\pm}$1.36 mm (F-4), 12.4${\pm}$1.41 mm (F-3),15.1${\pm}$1.57 mm (F-2), 18.1${\pm}$1.67 mm (F-1) and 21.2${\pm}$2.25 mm (F-0). In the BBT coverline endopint; 9.9${\pm}$.39 mm (C-4), 12.5 ${\pm}$1.44 mm (C-3), 15.2${\pm}$1.64 mm (C-2), 18.0 ${\pm}$1.69 mm (C-1), and 21.2${\pm}$2.31 mm (C-0). 5. The relationship between the ultrasonographic signs of ovulation and the basal body temperature curve were as following. The BBT dip correlated with the ovulation in 2 cases, which revealed decrease in follicular diameter (100%), fluid pattem in the Cul-de-Sac (1 case, 50%) and complete disappearance of follicle (1 case, 50%). In the nadir (5 cases); the ultrasonographic signs of ovulation were decrease in follicular diameter (5 cases, 100%), fluid pattern in the Cul-de-Sac (1 case, 20%) and complete disappearance of follicle (3 cases, 60%). In the First day of Hyperthermic Plateau (41 cases); decrease in follicular diameter (40 cases, 97.6%), fluid pattern in the Cul-de-Sac (11 cases, 26.8%), appearance of internal echo and thickening of the wall (6 cases, 14.6%) and com plete disappearance of follicle (28 cases, 68.3%). In the BBT coverline endpoint (35 cases); decrease in follicular diameter (33 cases, 94.3%), fluid pattern in the Cul-de Sac (9 cases, 25.7%), appearance of internal echo and thickening of the wall (5 cases 14.3%) and complete disappearance of follicle (20 cases, 57.1%).

• Journal of Veterinary Clinics
• /
• v.31 no.5
• /
• pp.394-402
• /
• 2014

The objective of this study was to investigate the effect of chronic unpredictable stress on the reproductive function and ovarian luteinizing hormone receptor (LHR) expression. 9-week-old C57BL/6 female mice were randomly divided into two groups: control group and stressed group. Mice have been stressed twice a day for 35 days with 12 different stressors which were randomly selected. The results demonstrate that there is significant increase in the anxiety-related behaviors (P < 0.05), decrease body weight gain rate (P < 0.01) and decrease in the average of litter size in stressed mice compared with control group (P < 0.01). Furthermore, the rate of primary, secondary and early antral follicles in stressed mice significantly decreased (P < 0.05), whereas that of atretic follicles significantly increased compared with control mice (P < 0.01). The immunohistochemical analysis revealed that reduced LHR expression in granulosa cells of follicle and luteal cells of corpus luteum in response to chronic unpredictable stress. The western blot analysis revealed significantly decrease in LHR expression in the stressed mice ovaries compared with the control (P < 0.05). These results suggest that ovarian LHR expression affected by chronic unpredictable stress and the modulated ovarian LHR is responsible for ovarian follicular maldevelopment and reproductive dysfunction.

• Korean Journal of Animal Reproduction
• /
• v.25 no.2
• /
• pp.119-124
• /
• 2001

This study was performed to investigate the effects of the peptide to carrier ratio on the immune and biological functions to inhibin immunization in Hanwoo. A peptide sequence kom the alpha -subunit (19~32 peptide) of porcine inhibin was synthesized for antigen and conjugated to human serum albumin(HSA) for carrier protein. Anti-inhibin sera(AI) were produced 52 day later from rabbit after injection of inhibin-$\alpha$ -subunit peptide conjugator for antigen with the interval of 2 weeks. Immune-blotting analysis using antibody specific fur inhibin-$\alpha$ subunits revealed that the inhibin was detected at 1.0 cm bovine follicular fluid(bFF). However, each stage of corpus lutea and 0.1 cm of follicular fluid were not detected. The maximal contents of estradiol-17 $\beta$ in Hanwoo ovarian follicular fluid were detected at 2.0 cm of follicular size(diameter), but the mean total contents of these hormone decreased significantly with decreasing diameter of follicles. However, progesterone contents of follicular fluid were high at 1.0 cm of follicle. Progesterone secretion by Hanwoo granulosa cell cultured for 48 hr in vitro was significantly (p＜0.05) inhibited in 5% bFF and 5% bFF + 5% AI addition group compared with control group. Estradiol-17 $\beta$ secretion by Hanwoo granulosa cell cultured for 48 hr in vitro was significantly (p＜0.05) increased in 5% AI and 5% AI + 5% bFF addtion group compared with control group. However, the groups added 5% AI were not changed compared to control groups in progesterone and estradiol-17 $\beta$. Taken together, we suggested that inhibin in the mature FF plays a pivotal role on the biosynthesis of steroid hormone of follicular cells during follicular development.

• Korean Journal of Animal Reproduction
• /
• v.6 no.1
• /
• pp.19-30
• /
• 1982

1. The cows slaughtered at age of 3, 4, 6, 7, 8, and 9 years old were 1.5, 1.5, 15.0, 62.5 and 4.4% respectively. 2. The cows slaughtered at 351-450kg and more than 500kg were 60 and 28% respectively. 3. Best, very good, good and bad cows in nutritional condition were 1.6, 25.8, 62.9, and 9.7% respectively. Among the six cows which were bad nutrition, the two were with severe endometritis, the three were normal in genital function and one was on 70 days of pregnancy. 4. Holstein cows(55.2%) showed higher reproductive failure than the Korean cows(33.3%). 5. The slaughted ratio of the Korean cattle and Holstein cows was 36 and 64% respectively. 6. Pregnant cows were about 16% among the slaughtered one. 7. Reproductive failures were composed of 46% in uterus, 32% in ovaries, 8% in udder, 6% in oviduct, 4% in cervix of uterine, 2% in vagina and 2% inmummified fetus. 8. Forty six percentages of uterine diseases were as follows; horn, 13%, body of uterus, 32% and ovary diseases were 32%, that is, 12% of ovary atrophy, 8% of ovarycyst and 6% of lutealcyst. 9. The cows of reproductive failures were commonly infected with 1.6 kinds of diseases. 10. According to classification, six type of ovaries were as follows; normal, 58%, ovary-cyst, 11%, luteum cyst, 4%, coexistence of follicles and corpus luteum, 16%, weak function of ovaries, 10% and ovarian atrophy, 1%. 11. Major axis, minor axis and thickness of right ovary were larger than those of left one both in Korean cattle and Holstein cows. Holstein cow had generally larger size of ovary than these of the Korean cattle.. 12. The left and right oviducts showed no difference in length, but Holstein had longer oviduct than Korean cow. 13. There was no difference in the length of uterine horn between right and left in the Korean cows, but the right was longer than the left in Holstein cows. 14. Holstein had longer horn and body of uterine than the Korean cows. 15. The weight of right ovary was heavier than that of left in both breeds, but there was no differences in weight of left ovary between two breeds and right ovary of Holstein breed was heavier than that of the Korean cow. 16. The weight of right oviduct and uterine born was heavier than that of the left, and Holstein had heavier oviducts and uterine horns than the Korean cows. 17. Holstein had heavier uterine body and cervix of uterine than the Korean cows. 18. The length of reproductive systems of Korean cow is as follows; Major and minor diameter and thickness ofovary are 3.6${\pm}$0.7, 2.3${\pm}$0.4 and 1.6${\pm}$1.4 cm in left and 3.7${\pm}$0.6, 2.5${\pm}$0.5 and 1.8${\pm}$0.5 cm in right. Oviduct is 28.4${\pm}$3.1 cm in left and 27.8${\pm}$3.3 cm in right. Uterine horn is 27.4${\pm}$4.5 cm in left and 27.7${\pm}$4.9 cm in right. Uterine body and cervix are 3.4${\pm}$1.1 and 6.5${\pm}$1.7 cm. 19. The length of female reproductive systems ofHolstein cow is as follows; Major and minor diameter and thickness of ovary are 3.9${\pm}$1.3, 2.3${\pm}$0.5, and 1.5${\pm}$0.6 cm in left and 4.0${\pm}$0.8, 2.8${\pm}$0.6 and 1.8${\pm}$0.6 cm in right. Oviduct is 29.4${\pm}$4.2 cm in left and 29.3${\pm}$4.1 cm in right. Uterine horn is 30.2${\pm}$7.4 cm in left and 32.6${\pm}$8.4 cm in right. Uterine body and cervix are 4.5${\pm}$2.5 and 7.8${\pm}$2.9 cm. 20. The weight of reproductive systems of Korean cow is as follows; Ovary is 8.4${\pm}$4.1 g in left and 9.3${\pm}$3.6g in right. Oviduct is 1.5${\pm}$0.5 g in left and 1.6${\pm}$0.5 g in right. Uterine horn is 109${\pm}$27 g left and 118${\pm}$32 g in right. Uterine body and cervix are 30.4${\pm}$14.1 and 76.7${\pm}$38.4g. 21. The weight of reproductive systems of Holstein cow is as follows; Ovary is 8.2${\pm}$3.1 g in left and 12.5${\pm}$5.6 g in right. Oviduct is 1.7${\pm}$0.6 g in left and 1.9${\pm}$0.9 g in right. Uterine horn is 199${\pm}$14.2 g in left and 221${\pm}$111.2g in right. Uterine body and cervix are 58.2${\pm}$46.5 and 126.7${\pm}$103.3 g.