• 스마트미디어저널
• /
• 제10권1호
• /
• pp.39-46
• /
• 2021

본 연구는 뉴스 기사에서 기사문을 구성하는 문장별 중요도를 고려하여 요약문을 추출하는 방법에 관한 것으로 문장 중요도에 영향을 주는 특성으로 중심 문장(Topic Sentence)일 확률, 기사 제목 및 다른 문장과의 유사도, 문장 위치에 따른 가중치를 추출하여 문장 중요도를 계산하는 방법을 제안한다. 이때, 중심 문장(Topic Sentence)은 일반 문장과는 구별되는 특징을 가질 것이라는 가설을 세우고, 딥러닝 기반 분류 모델을 학습시켜 입력 문장에 대한 중심 문장 확률값을 구한다. 또한 사전학습된 ELMo 언어 모델을 활용하여 문맥 정보를 반영한 문장 벡터값을 기준으로 문장간 유사도를 계산하여 문장 특성으로 추출한다. LSTM 및 BERT 모델의 중심 문장 분류성능은 정확도 93%, 재현율 96.22%, 정밀도 89.5%로 높은 분석 결과가 나왔으며, 이렇게 추출된 문장 특성을 결합하여 문장별 중요도를 계산한 결과, 기존 TextRank 알고리즘과 비교하여 중심 문장 추출 성능이 10% 정도 개선된 것을 확인할 수 있었다.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제35권4호
• /
• pp.255-264
• /
• 2008

본 논문은 K-means과 비음수 행렬 분해(NMF)를 이용하여 주제기반의 다중문서를 요약하는 새로운 방법을 제안하였다. 제안방법은 비음수 행렬 분해를 이용하여 가중치가 부여된 용어-문장 행렬을 희소(Sparse)한 비음수 의미특징 행렬과 비음수 변수 행렬로 분해함으로써 직관적으로 이해할 수 있는 형태의 의미적 특징을 추출할 수 있고, 주제와 의미특징간의 유사도에 가중치를 부여하여 유사도는 높으나 실제 의미 없는 문장이 추출되는 것을 막는다. 또한 K-means 군집을 이용하여 문장에 포함된 노이즈를 제거함으로써 문서의 의미가 요약에 편향되게 반영하는 것을 피할 수 있고, 추출된 문장에 부여된 순위순서대로 정렬하여 보여 줌으로써 응집성을 높인다. 실험 결과 제안방법이 다른 방법에 비하여 좋은 성능을 보인다.

• Journal of Information Processing Systems
• /
• 제18권6호
• /
• pp.794-802
• /
• 2022

The recovery of reasonable depth information from different scenes is a popular topic in the field of computer vision. For generating depth maps with better details, we present an efficacious monocular depth prediction framework with coordinate attention and feature fusion. Specifically, the proposed framework contains attention, multi-scale and feature fusion modules. The attention module improves features based on coordinate attention to enhance the predicted effect, whereas the multi-scale module integrates useful low- and high-level contextual features with higher resolution. Moreover, we developed a feature fusion module to combine the heterogeneous features to generate high-quality depth outputs. We also designed a hybrid loss function that measures prediction errors from the perspective of depth and scale-invariant gradients, which contribute to preserving rich details. We conducted the experiments on public RGBD datasets, and the evaluation results show that the proposed scheme can considerably enhance the accuracy of depth prediction, achieving 0.051 for log10 and 0.992 for δ<1.25³ on the NYUv2 dataset.

• Journal of Information Processing Systems
• /
• 제17권1호
• /
• pp.163-177
• /
• 2021

Many methods of discovering social networking communities or clustering of features are based on the network structure or the content network. This paper proposes a community discovery method based on topic models using a time factor and an unsupervised clustering method. Online community discovery enables organizations and businesses to thoroughly understand the trend in users' interests in their products and services. In addition, an insight into customer experience on social networks is a tremendous competitive advantage in this era of ecommerce and Internet development. The objective of this work is to find clusters (communities) such that each cluster's nodes contain topics and individuals having similarities in the attribute space. In terms of social media analytics, the method seeks communities whose members have similar features. The method is experimented with and evaluated using a Vietnamese corpus of comments and messages collected on social networks and ecommerce sites in various sectors from 2016 to 2019. The experimental results demonstrate the effectiveness of the proposed method over other methods.

• 대한음성학회지:말소리
• /
• 제29_30호
• /
• pp.19-42
• /
• 1995

This paper deals with the current phonological theory named the elementary theory Before manipulating this theory with highest depth, we need to discuss the distinctive feature theory which can be estimated as playing a central role in the standard generative trends. En the element-based phonological theory, the followings might act as tile main traits in differentiating it from other theories: the notion of phonological opposition is regarded as privative one, treating the univalent element as an analysing unit in a phonological process instead of distinctive features, and the nile convention in standard theory is replaced by the element. In chapter 2, a brief history of generative theory is to be described with respect to the merits and demerits of the distinctive feature theory. In chapter 3, After dealing with the current tendency and some prominent aspects of each element theory, a couple of problems thought to be confronted by the distinctive feature theory And the analysing method taken from the element-based theory which may be regarded currently as the alternative to the problems mentioned above will be discussed mainly, The government-based theory introduced by KLV(1985, 1988) may be the main target in discussing the current topic.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권11호
• /
• pp.5592-5609
• /
• 2017

JPEG steganography detection is an active research topic in the field of information hiding due to the wide use of JPEG image in social network, image-sharing websites, and Internet communication, etc. In this paper, a new steganalysis method for content-adaptive JPEG steganography is proposed by integrating the evolutionary feature selection and classifier ensemble selection. First, the whole framework of the proposed steganalysis method is presented and then the characteristic of the proposed method is analyzed. Second, the feature selection method based on genetic algorithm is given and the implement process is described in detail. Third, the method of classifier ensemble selection is proposed based on Pareto evolutionary optimization. The experimental results indicate the proposed steganalysis method can achieve a competitive detection performance by compared with the state-of-the-art steganalysis methods when used for the detection of the latest content-adaptive JPEG steganography algorithms.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권1호
• /
• pp.131-146
• /
• 2021

Multimodal biometric-based recognition has been an active topic because of its higher convenience in recent years. Due to high user convenience of finger, finger-based personal identification has been widely used in practice. Hence, taking Finger-Print (FP), Finger-Vein (FV) and Finger-Knuckle-Print (FKP) as the ingredients of characteristic, their feature representation were helpful for improving the universality and reliability in identification. To usefully fuse the multimodal finger-features together, a new robust representation algorithm was proposed based on hierarchical model. Firstly, to obtain more robust features, the feature maps were obtained by Gabor magnitude feature coding and then described by Local Binary Pattern (LBP). Secondly, the LGBP-based feature maps were processed hierarchically in bottom-up mode by variable rectangle and circle granules, respectively. Finally, the intension of each granule was represented by Local-invariant Gray Features (LGFs) and called Hierarchical Local-Gabor-based Gray Invariant Features (HLGGIFs). Experiment results revealed that the proposed algorithm is capable of improving rotation variation of finger-pose, and achieving lower Equal Error Rate (EER) in our homemade database.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권11호
• /
• pp.4395-4412
• /
• 2020

The multisource image fusion has become an active topic in the last few years owing to its higher segmentation rate. To enhance the accuracy of multimodal pig-body feature segmentation, a multisource image fusion method was employed. Nevertheless, the conventional multisource image fusion methods can not extract superior contrast and abundant details of fused image. To superior segment shape feature and detect temperature feature, a new multisource image fusion method was presented and entitled as NSST-GF-IPCNN. Firstly, the multisource images were resolved into a range of multiscale and multidirectional subbands by Nonsubsampled Shearlet Transform (NSST). Then, to superior describe fine-scale texture and edge information, even-symmetrical Gabor filter and Improved Pulse Coupled Neural Network (IPCNN) were used to fuse low and high-frequency subbands, respectively. Next, the fused coefficients were reconstructed into a fusion image using inverse NSST. Finally, the shape feature was extracted using automatic threshold algorithm and optimized using morphological operation. Nevertheless, the highest temperature of pig-body was gained in view of segmentation results. Experiments revealed that the presented fusion algorithm was able to realize 2.102-4.066% higher average accuracy rate than the traditional algorithms and also enhanced efficiency.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제6권10호
• /
• pp.2632-2649
• /
• 2012

Most approaches to human action recognition is limited due to the use of simple action datasets under controlled environments or focus on excessively localized features without sufficiently exploring the spatio-temporal information. This paper proposed a framework for recognizing realistic human actions. Specifically, a new action representation is proposed based on computing a rich set of descriptors from keypoint trajectories. To obtain efficient and compact representations for actions, we develop a feature fusion method to combine spatial-temporal local motion descriptors by the movement of the camera which is detected by the distribution of spatio-temporal interest points in the clips. A new topic model called Markov Semantic Model is proposed for semantic feature selection which relies on the different kinds of dependencies between words produced by "syntactic " and "semantic" constraints. The informative features are selected collaboratively based on the different types of dependencies between words produced by short range and long range constraints. Building on the nonlinear SVMs, we validate this proposed hierarchical framework on several realistic action datasets.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권2호
• /
• pp.604-625
• /
• 2018

Nowadays, SMS spam has been overflowing in many countries. In fact, the standards of filtering SMS spam are different from country to country. However, the current technologies and researches about SMS spam filtering all focus on dividing SMS message into two classes: legitimate and illegitimate. It does not conform to the actual situation and need. Furthermore, they are facing several difficulties, such as: (1) High quality and large-scale SMS spam corpus is very scarce, fine categorized SMS spam corpus is even none at all. This seriously handicaps the researchers' studies. (2) The limited length of SMS messages lead to lack of enough features. These factors seriously degrade the performance of the traditional classifiers (such as SVM, K-NN, and Bayes). In this paper, we present a new fine categorized SMS spam corpus which is unique and the largest one as far as we know. In addition, we propose a classifier, which is based on the probability topic model. The classifier can alleviate feature sparse problem in the task of SMS spam filtering. Moreover, we compare the approach with three typical classifiers on the new SMS spam corpus. The experimental results show that the proposed approach is more effective for the task of SMS spam filtering.