• Journal of Internet Computing and Services
• /
• v.16 no.1
• /
• pp.29-37
• /
• 2015

In this paper, we propose a robust pre-processing module for camera-based Optical Music Score Recognition (OMR) on mobile device. The captured images likely suffer for recognition from many distortions such as illumination, blur, low resolution, etc. Especially, the complex background music sheets recognition are difficult. Through any symbol recognition system, the staff line height and staff line space are used many times and have a big impact on recognition module. A robust and accurate staff line height and staff line space are essential. Some staff line height and staff line space are proposed for binary image. But in case of complex background music sheet image, the binarization results from common binarization algorithm are not satisfactory. It can cause incorrect staff line height and staff line space estimation. We propose a robust staff line height and staff line space estimation by using run-length encoding technique on edge image. Proposed method is composed of two steps, first step, we conducted the staff line height and staff line space estimation based on edge image using by Sobel operator on image blocks. Each column of edge image is encoded by run-length encoding algorithm Second step, we detect the staff line using by Stable Path algorithm and removal the staff line using by adaptive Line Track Height algorithm which is to track the staff lines positions. The result has shown that robust and accurate estimation is possible even in complex background cases.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.303-310
• /
• 2007

Named Entity Recognition (NER) is always limited by its lower recall resulting from the asymmetric data distribution where the NONE class dominates the entity classes. This paper presents an approach that exploits non-local information to improve the NER recall. Several kinds of non-local features encoding entity token occurrence, entity boundary and entity class are explored under Conditional Random Fields (CRFs) framework. Experiments on SIGHAN 2006 MSRA (CityU) corpus indicate that non-local features can effectively enhance the recall of the state-of-the-art NER systems. Incorporating the non-local features into the NER systems using local features alone, our best system achieves a 23.56% (25.26%) relative error reduction on the recall and 17.10% (11.36%) relative error reduction on the F1 score; the improved F1 score 89.38% (90.09%) is significantly superior to the best NER system with F1 of 86.51% (89.03%) participated in the closed track.

• Annual Conference on Human and Language Technology
• /
• 2008.10a
• /
• pp.156-161
• /
• 2008

This work presents an agenda-based approach to improve the robustness of the dialog manager by using dialog examples and n-best recognition hypotheses. This approach supports n-best hypotheses in the dialog manager and keeps track of the dialog state using a discourse interpretation algorithm with the agenda graph and focus stack. Given the agenda graph and n-best hypotheses, the system can predict the next system actions to maximize multi-level score functions. To evaluate the proposed method, a spoken dialog system for a building guidance robot was developed. Preliminary evaluation shows this approach would be effective to improve the robustness of example-based dialog modeling.

• Clinics in Shoulder and Elbow
• /
• v.23 no.2
• /
• pp.62-70
• /
• 2020

Background: Recent studies have reported high rates of recurrence of shoulder instability in patients with glenoid bone defects greater than 20% after capsulolabral reconstruction. The purpose of the present study was to evaluate the failure rate of arthroscopic capsulolabral reconstruction for the treatment of anterior instability in the presence of glenoid bone deficits >20%. Methods: Retrospective analyses were conducted among cases with anterior shoulder instability and glenoid bone defects of >20% that were treated by arthroscopic capsulolabral reconstruction with a minimum 2-year follow-up (30 cases). We included the following variables: age, bone defect size, instability severity index score (ISIS), on-/off-track assessment, incidence recurrent instability, and return to sports. Results: The mean glenoid bone defect size was 25.8%±4.2% (range, 20.4%-37.2%), and 18 cases (60%) had defects of >25%. Bony Bankart lesions were identified in 11 cases (36.7%). Eleven cases (36.7%) had ISIS scores >6 points and 21 cases (70%) had off-track lesions. No cases of recurrent instability were identified over a mean follow-up of 39.9 months (range, 24-86 months), but a sense of subluxation was reported by three patients. Return to sports at the preinjury level was possible in 24 cases (80%), and the average satisfaction rating was 92%. Conclusions: Arthroscopic soft tissue reconstruction was successful for treating anterior shoulder instability among patients with glenoid bone defects >20%, even enabling return to sports. Future studies should focus on determining the range of bone defect sizes that can be successfully managed by soft tissue repair.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.5
• /
• pp.221-226
• /
• 2024

In information retrieval, queries come in various types, ranging from abstract queries to those containing specific keywords, making it a challenging task to accurately produce results according to user demands. Additionally, search systems must handle queries encompassing various elements such as typos, multilingualism, and codes. Reranking is performed through training suitable documents for queries using DeBERTa, a deep learning model that has shown high performance in recent research. To evaluate the effectiveness of the proposed method, experiments were conducted using the test collection of the Product Search Track at the TREC 2023 international information retrieval evaluation competition. In the comparison of NDCG performance measurements regarding the experimental results, the proposed method showed a 10.48% improvement over BM25, a basic information retrieval model, in terms of search through query error handling, provisional relevance feedback-based product title-based query expansion, and reranking according to query types, achieving a score of 0.7810.

• Journal of Science Education
• /
• v.42 no.3
• /
• pp.308-321
• /
• 2018

The purpose of the study is to develop a questionnaire that examines unskeptical attitudes in scientific inquiry context. The questionnaire items were developed through literature research, expert review, and statistical analyses for validity and the differences in scores were identified by gender and tracks. A total of 363 high school students participated in the study. To explore the validity evidence of items, the Rasch analysis and the reliability of internal consistency were performed, and the two-way ANOVA was performed to compare the scores of the unskeptical attitudes between gender and academic track. Self-reporting and Likert-scaling 23 items were developed to measure unskeptical attitudes in scientific inquiry context. The items were developed in the sub-domain of scientific inquiry: 'questioning and hypothesis generating,' 'experiment designing,' and 'explaining and interpreting.' Second, the validity and reliability of the unskeptical were identified in a rigorous method. The validity of items were identified by multi-dimensional partial score model analysis through the Rasch model, and all 23 items were found to be fit to model. Various reliability evidences were also found to be appropriate. It was found that there were no significant differences of unskeptical attitude score between the gender and academic track except one comparison. The developed questionnaire could be used to check an unskeptical attitude in the course of scientific inquiry and to compare the effects of scientific inquiry classes.

• Journal of Chest Surgery
• /
• v.51 no.4
• /
• pp.247-253
• /
• 2018

Background: Early extubation after cardiovascular surgery has some clinical advantages, including reduced hospitalization costs. Herein, we review the results of ultra-fast-track (UFT) extubation, which refers to extubation performed on the operating table just after the operation, or within 1-2 hours after surgery, in patients with congenital cardiac disease. Methods: We performed UFT extubation in patients (n=72) with a relatively simple congenital cardiac defect or who underwent a simple operation starting in September 2016. To evaluate the feasibility and effectiveness of our recently introduced UFT extubation strategy, we retrospectively reviewed 195 patients who underwent similar operations for similar diseases from September 2015 to September 2017, including the 1-year periods immediately before and after the introduction of the UFT extubation protocol. Propensity scores were used to assess the effects of UFT extubation on length of stay (LOS) in the intensive care unit (ICU), hospital LOS, and medical costs. Results: After propensity-score matching using logistic regression analysis, 47 patients were matched in each group. The mean ICU LOS ($16.3{\pm}28.6$ [UFT] vs. $28.0{\pm}16.8$ [non-UFT] hours, p=0.018) was significantly shorter in the UFT group. The total medical costs ($182.6{\pm}3.5$ [UFT] vs. $187.1{\pm}55.6$ [non-UFT] ${\times}100,000$ Korean won [KRW], p=0.639) and hospital stay expenses ($48.3{\pm}13.6$ [UFT] vs. $54.8{\pm}29.0$ [non-UFT] ${\times}100,000KRW$, p=0.164) did not significantly differ between the groups. Conclusion: UFT extubation decreased the ICU LOS and mechanical ventilation time, but was not associated with postoperative hospital LOS or medical expenses in patients with simple congenital cardiac disease.

• Journal of the Society of Naval Architects of Korea
• /
• v.57 no.3
• /
• pp.140-151
• /
• 2020

It is important to identify the location of ship hull blocks with exact block identification number when scheduling the shipbuilding process. The wrong information on the location and identification number of some hull block can cause low productivity by spending time to find where the exact hull block is. In order to solve this problem, it is necessary to equip the system to track the location of the blocks and to identify the identification numbers of the blocks automatically. There were a lot of researches of location tracking system for the hull blocks on the stockyard. However there has been no research to identify the hull blocks on the stockyard. This study compares the performance of 5 Convolutional Neural Network (CNN) models with multi-view image set on the classification of the hull blocks to identify the blocks on the stockyard. The CNN models are open algorithms of ImageNet Large-Scale Visual Recognition Competition (ILSVRC). Four scaled hull block models are used to acquire the images of ship hull blocks. Learning and transfer learning of the CNN models with original training data and augmented data of the original training data were done. 20 tests and predictions in consideration of five CNN models and four cases of training conditions are performed. In order to compare the classification performance of the CNN models, accuracy and average F1-Score from confusion matrix are adopted as the performance measures. As a result of the comparison, Resnet-152v2 model shows the highest accuracy and average F1-Score with full block prediction image set and with cropped block prediction image set.

• Journal of the Korean earth science society
• /
• v.22 no.6
• /
• pp.469-479
• /
• 2001

This study investigated the trends in the number of applicants and mean score and applicants’ responses on the test items of Earth Science in the College Scholastic Ability Test (CSAT) implemented for 3 years (1999-2001). The percentage of applicants of science track were 43.14% in 1995, but reduced through 29.5% in 2001 to 26.92% in 2002 CSAT For elective subject, about 22% of science track students applied to Earth Science II which ranked third to Biology II and Chemistry II. In 1999, test items were developed to have the expected difficulty 40 ${\sim}$ 59% (6 items) to 60 ${\sim}$ 79% (10 items). But in 2001 every 16 items were developed to have difficulty 60 ${\sim}$ 79%, which was caused by the policy of so called ‘easy CSAT’. Thus the mean score of ‘Earth Science II’ was increased from 50.26 in 1999 through 64.47 in 2000, to 67.58 in 2001. Applicants were generally very good at solving test items focusing on process skills only and familar items but poor at solving test items related to the motion of the earth and planets and sea wave, especially items calling two or more concepts. Thus special measures to cope with the decrease in applicants of science track should be provided. And it is recommended to develop test items with wider range of difficulty and to reduce test items calling process skills only. And special consideration should be given to teaching the content area with poor achievement and high actual difficulty compared to the expected.

• Journal of Science Education
• /
• v.39 no.1
• /
• pp.99-112
• /
• 2015

This study aimed at investigating elementary teachers' stages of concerns and its changes about STEAM education. The participants of this study were 90 elementary school teachers implementing STEAM education in their schools. The Stages of Concern Questionnaire (SoCQ) was administered after the permission was granted by Hall. Data were collected three times from April, June, and December in the year of implementing the exemplary STEAM school. The results indicated that elementary school teachers' concerns toward STEAM education, the differences of the percentile scores of each stages of concerns, showed slightly low. The SoC of teachers in April showed that Awareness(Stage 0) was relatively very high and Consequences(Stage 4) was very low. However, in December, both Awareness(Stage 0) and Management(Stage 3) were very low. In particular, the percentile scores of Consequence(Stage 4) showed the big difference from 54.7 to 74.3. In addition, teachers who took science track in their high schools showed the relatively low score in the stage of Management(Stage 3). Teachers having the degree of Master of Arts presented the relatively low percentile score in the stage of Management(Stage 3). Teachers who majored in STEM related areas at the university presented the relatively low score in the stage of Management(Stage 3). The findings of this exploratory study may provide the useful insights into the integrative approaches of STEAM education.