• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2018년도 제30회 한글 및 한국어 정보처리 학술대회
• /
• pp.259-262
• /
• 2018

기계독해의 목표는 기계가 주어진 문맥을 이해하고 문맥에 대한 질문에 대답할 수 있도록 하는 것이다. 본 논문에서는 Multi-level Attention에 정보를 효율적으로 융합 수 있는 Fusion 함수를 결합하고, Answer module에Stochastic multi-step answer를 적용하여 SQuAD dev 데이터 셋에서 EM=78.63%, F1=86.36%의 성능을 보였다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제14권4호
• /
• pp.332-339
• /
• 2014

Face classification has wide applications in security and surveillance. However, this technique presents various challenges caused by pose, illumination, and expression changes. Face recognition with long-distance images involves additional challenges, owing to focusing problems and motion blurring. Multiple frames under varying spatial or temporal settings can acquire additional information, which can be used to achieve improved classification performance. This study investigates the effectiveness of multi-frame decision-level fusion with photon-counting linear discriminant analysis. Multiple frames generate multiple scores for each class. The fusion process comprises three stages: score normalization, score validation, and score combination. Candidate scores are selected during the score validation process, after the scores are normalized. The score validation process removes bad scores that can degrade the final output. The selected candidate scores are combined using one of the following fusion rules: maximum, averaging, and majority voting. Degraded facial images are employed to demonstrate the robustness of multi-frame decision-level fusion in harsh environments. Out-of-focus and motion blurring point-spread functions are applied to the test images, to simulate long-distance acquisition. Experimental results with three facial data sets indicate the efficiency of the proposed decision-level fusion scheme.

• Journal of Information Processing Systems
• /
• 제15권5호
• /
• pp.1243-1257
• /
• 2019

As current algorithms unable to perform effective fusion processing of unknown complex radar signals lacking database, and the result is unstable, this paper presents a multi-level fusion processing algorithm for complex radar signals based on evidence theory as a solution to this problem. Specifically, the real-time database is initially established, accompanied by similarity model based on parameter type, and then similarity matrix is calculated. D-S evidence theory is subsequently applied to exercise fusion processing on the similarity of parameters concerning each signal and the trust value concerning target framework of each signal in order. The signals are ultimately combined and perfected. The results of simulation experiment reveal that the proposed algorithm can exert favorable effect on the fusion of unknown complex radar signals, with higher efficiency and less time, maintaining stable processing even of considerable samples.

• 정보보호학회논문지
• /
• 제29권5호
• /
• pp.1007-1017
• /
• 2019

현재 대부분의 생체인증 시스템은 단일 생체정보를 이용하여 사용자를 인증하고 있는데, 이러한 방식은 노이즈로 인한 문제, 데이터에 대한 민감성 문제, 스푸핑, 인식률의 한계 등 많은 문제점들을 가지고 있다. 이를 해결하기 위한 방법 중 하나로 다중 생체정보를 이용하는 방법이 제시되고 있다. 다중 생체인증 시스템은 각각의 생체정보에 대해서 information fusion을 수행하여 새로운 정보를 생성한 뒤, 그 정보를 활용하여 사용자를 인증하는 방식이다. Information fusion 방법들 중에서 score-level fusion 방법을 보편적으로 많이 사용한다. 하지만 정규화 작업이 필요하다는 문제점을 갖고 있고, 데이터가 같아도 정규화 방법에 따라 인식률이 달라진다는 문제점을 갖고 있다. 이에 대한 대안으로 정규화 작업이 필요 없는 rank-level fusion 방법이 제시되고 있다. 하지만 기존의 rank-level fusion 방법들은 score-level fusion 방법보다 인식률이 낮다. 이러한 문제점을 해결하기 위해 상관계수를 이용하여 score-level fusion 방법보다 인식률이 높은 rank-level fusion 방법을 제안한다. 실험은 홍채정보(CASIA V3)와 얼굴정보(FERET V1)를 이용하여 기존의 존재하는 rank-level fusion 방법들의 인식률과 본 논문에서 제안하는 fusion 방법의 인식률을 비교하였다. 또한 score-level fusion 방법들과도 인식률을 비교하였다. 그 결과로 인식률이 약 0.3%에서 3.3%까지 향상되었다.

• Journal of Korean Neurosurgical Society
• /
• 제30권sup1호
• /
• pp.55-60
• /
• 2001

Objective : The purpose of the present study was to examine neurologic changes, fusion rate and degree of kyphosis from the surgical results of those patients who underwent multi-level anterior interbody fusion and internal fixation. Methods : Among 63 cases of the patients who received multi-level anterior interbody fusion and internal fixation in 5 years between 1995 to 1999 at the neurosurgery department, we performed a retrospective study in 52 cases that could be followed up with dynamic view imaging ; the results were compared and analyzed. The analysis was based on the results of history taking, physical findings and radiologic findings, and Odom criteria were used to classify those cases with neurologic changes. Results : Among those 52 cases in whom the follow-up was possible for at least a year and dynamic view imaging was possible, bone fusion was seen in 93% of the trauma cases and 95% in the non-trauma cases and overall bone fusion was observed in 94% of the cases. Bone fusion was seen in 93% of the autobone cases, 95% of the allobone cases, and 94% of the Mesh cases. Radiologic changes were observed by comparing the lateral view after surgery ; kyphosis was seen in 53% of the autobone cases, in 70% of the allobone cases, and in 35% of Mesh cases ; in 45% and 60% of the non-trauma cases and trauma cases, respectively ; and in 55% of the 2 level fusion cases and 46% of the 3 level fusion cases. Neurologic changes classified according to Odom criteria showed excellent result in 48% of all the cases, good in 23%, fair in 4%, and poor in 25%. Conclusion : Even those cases with multi-level fusion, a high fusion rate could be obtained by performing anterior interbody fusion and internal fixation ; those cases with kyphosis were related more with the presence or absence of posterior compartment injury rather than the fusion level ; and those trauma cases showed not much difference in the fusion rate compared with non-trauma cases but had a high possibility of kyphosis.

• Journal of Information Processing Systems
• /
• 제14권1호
• /
• pp.176-190
• /
• 2018

This paper presents a novel convolutional neural network based multi-feature fusion learning method for non-rigid 3D model retrieval, which can investigate the useful discriminative information of the heat kernel signature (HKS) descriptor and the wave kernel signature (WKS) descriptor. At first, we compute the 2D shape distributions of the two kinds of descriptors to represent the 3D model and use them as the input to the networks. Then we construct two convolutional neural networks for the HKS distribution and the WKS distribution separately, and use the multi-feature fusion layer to connect them. The fusion layer not only can exploit more discriminative characteristics of the two descriptors, but also can complement the correlated information between the two kinds of descriptors. Furthermore, to further improve the performance of the description ability, the cross-connected layer is built to combine the low-level features with high-level features. Extensive experiments have validated the effectiveness of the designed multi-feature fusion learning method.

• Journal of Information Processing Systems
• /
• 제10권4호
• /
• pp.555-567
• /
• 2014

Biometric performance improvement is a challenging task. In this paper, a hierarchical strategy fusion based on multimodal biometric system is presented. This strategy relies on a combination of several biometric traits using a multi-level biometric fusion hierarchy. The multi-level biometric fusion includes a pre-classification fusion with optimal feature selection and a post-classification fusion that is based on the similarity of the maximum of matching scores. The proposed solution enhances biometric recognition performances based on suitable feature selection and reduction, such as principal component analysis (PCA) and linear discriminant analysis (LDA), as much as not all of the feature vectors components support the performance improvement degree.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2002년도 ITC-CSCC -2
• /
• pp.1244-1247
• /
• 2002

This paper discusses various algorithms to the fusion of multi-spectral image. These fusion techniques have a wide variety of applications that range from hospital pathology to battlefield management. Different algorithms in each fusion level, namely data, feature, and decision are compared. The PCT-Based algorithm, which has the characteristic of data compression, is described. The algorithm is experimented on a foliated aerial scene and the fusion result is presented.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권5호
• /
• pp.2253-2272
• /
• 2018

We propose a deep learning method for multi-focus image fusion. Unlike most existing pixel-level fusion methods, either in spatial domain or in transform domain, our method directly learns an end-to-end fully convolutional two-stream network. The framework maps a pair of different focus images to a clean version, with a chain of convolutional layers, fusion layer and deconvolutional layers. Our deep fusion model has advantages of efficiency and robustness, yet demonstrates state-of-art fusion quality. We explore different parameter settings to achieve trade-offs between performance and speed. Moreover, the experiment results on our training dataset show that our network can achieve good performance with subjective visual perception and objective assessment metrics.

• Journal of Korean Neurosurgical Society
• /
• 제50권5호
• /
• pp.441-445
• /
• 2011

Objective : There are technical limitations of multi-level posterior pedicle screw fixation performed by the percutaneous technique. The purpose of this study was to describe the surgical technique and outcome of minimally invasive multi-level posterior lumbar interbody fusion (PLIF) and to determine its efficacy. Methods : Forty-two patients who underwent mini-open PLIF using the percutaneous screw fixation system were studied. The mean age of the patients was 59.1 (range, 23 to 78 years). Two levels were involved in 32 cases and three levels in 10 cases. The clinical outcome was assessed using the visual analog scale (VAS) and Low Back Outcome Score (LBOS). Achievement of radiological fusion, intra-operative blood loss, the midline surgical scar and procedure related complications were also analyzed. Results : The mean follow-up period was 25.3 months. The mean LBOS prior to surgery was 34.5, which was improved to 49.1 at the final follow up. The mean pain score (VAS) prior to surgery was 7.5 and it was decreased to 2.9 at the last follow up. The mean estimated blood loss was 238 mL (140-350) for the two level procedures and 387 mL (278-458) for three levels. The midline surgical scar was 6.27 cm for two levels and 8.25 cm for three level procedures. Complications included two cases of asymptomatic medial penetration of the pedicle border. However, there were no signs of neurological deterioration or fusion failure. Conclusion : Multi-level, minimally invasive PLIF can be performed effectively using the percutaneous transpedicular screw fixation system. It can be an alternative to the traditional open procedures.