• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.5
• /
• pp.1671-1686
• /
• 2022

Face manipulation tools represented by Deepfake have threatened the security of people's biological identity information. Particularly, manipulation tools with deep learning technology have brought great challenges to Deepfake detection. There are many solutions for Deepfake detection based on traditional machine learning and advanced deep learning. However, those solutions of detectors almost have problems of poor performance when evaluated on different quality datasets. In this paper, for the sake of making high-quality Deepfake datasets, we provide a preprocessing method based on the image pixel matrix feature to eliminate similar images and the residual channel attention network (RCAN) to resize the scale of images. Significantly, we also describe a Deepfake detector named Cascaded-Hop which is based on the PixelHop++ system and the successive subspace learning (SSL) model. By feeding the preprocessed datasets, Cascaded-Hop achieves a good classification result on different manipulation types and multiple quality datasets. According to the experiment on FaceForensics++ and Celeb-DF, the AUC (area under curve) results of our proposed methods are comparable to the state-of-the-art models.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2000.11a
• /
• pp.45-52
• /
• 2000

Genomic approach produces massive amount of data within a short time period, New high-throughput automatic sequencers can generate over a million nucleotide sequence information overnight. A typical DNA chip experiment produces tens of thousands expression information, not to mention the tens of megabyte image files, These data must be handled automatically by computer and stored in electronic database, Thus there is a need for systematic approach of data collection, processing, and analysis. DNA sequence information is translated into amino acid sequence and is analyzed for key motif related to its biological and/or biochemical function. Functional genomics will play a significant role in identifying novel drug targets and diagnostic markers for serious diseases. As an enabling technology for functional genomics, bioinformatics is in great need worldwide, In Korea, a new functional genomics project has been recently launched and it focuses on identi☞ing genes associated with cancers prevalent in Korea, namely gastric and hepatic cancers, This involves gene discovery by high throughput sequencing of cancer cDNA libraries, gene expression profiling by DNA microarray and proteomics, and SNP profiling in Korea patient population, Our bioinformatics team will support all these activities by collecting, processing and analyzing these data.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2005.05a
• /
• pp.77-80
• /
• 2005

생물정보원은 이질성이 높고 사용자의 요구사항이 다양하다. 본 논문은 이러한 이질성을 해결하고 사용자의 다양한 요구사항에 쉽게 대처할 수 있는 XML 기반의 생물정보원 통합시스템의 설계개념과 구조 및 구현결과를 제시한다. 제시하는 통합시스템은 관계형테이블, 객체, XML, 플랫파일 등 다양한 자료형을 지원하면서, 관계형, 객체관계형, 웹자원, 응용프로그램 등 데이터 관리모델에 무관한 뷰 정의 및 질의처리모델이다. 그리고 사용자정의 XML 뷰 기반의 뷰 관리 및 질의처리를 통하여 사용자의 다양한 요구사항에 쉽게 적응할 수 있는 미디에이터 질의처리 기반의 생물정보원 통합시스템을 제시한다.

• Journal of Information Processing Systems
• /
• v.13 no.6
• /
• pp.1602-1612
• /
• 2017

Since rapidly disseminating of Internet of Things (IoT) as the new communication paradigm, a number of studies for various applications is being carried out. Especially, interest in the smart medical system is rising. In the smart medical system, a number of medical devices are distributed in popular area such as station and medical center, and this high density of medical device distribution can cause serious performance degradation of communication, referred to as the coexistence problem. When coexistence problem occurs in smart medical system, reliable transmitting of patient's biological information may not be guaranteed and patient's life can be jeopardized. Therefore, coexistence problem in smart medical system should be resolved. In this paper, we propose a distributed coexistence mitigation scheme for IoT-based smart medical system which can dynamically avoid interference in coexistence situation and can guarantee reliable communication. To evaluate the performance of the proposed scheme, we perform extensive simulations by comparing with IEEE 802.15.4 MAC protocol which is a traditional low-power communication technology.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.11a
• /
• pp.141-142
• /
• 2009

The PEFP (Proton Engineering Frontier Project) is constructing a 100MeV proton Linac (Linear Accelerator). The 20 MeV 20 mA proton beam has been serviced for an application in the fields of material, biological, information technology and medical sciences. For a stable and efficient acceleration of a proton beam, the control requirements must be optimized by studying various control methods. We propose that the integrated control system for the Linac machine components must be based on a distribution control method to improve a centralized control system. Based on EPICS (Experimental Physics and Industrial Control System) real-time software, the VME (Versa Module European package format) IOC (Input Output Controller) was developed under cross development environment with a RISC (Reduced Instruction Set Computer) PowerPC system. In this paper, we describe the design and implementation of distributed control system using the VME-based EPICS middleware for various components of the large proton accelerator.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.490-492
• /
• 2023

시퀀싱(sequencing) 기술의 발달로 다양한 오믹스(omics) 데이터의 축적과 인공 지능 기술의 발달로 인하여 다양한 드라이버 유전자 분류기법이 제안되어왔다. 최근에는 암 데이터가 대용량으로 축적되며 기계 학습 기반의 다양한 기법들이 활발히 제안되었다. 특히 다양한 오믹스 데이터를 결합한 고차원 데이터에서 높은 정확도를 확보하기 위한 시도가 활발히 이루어지고 있다. 본 논문에서는 멀티 오믹스와 네트워크 관련 특징을 기반으로 암의 증식 및 발생에 중요한 역할을 하는 드라이버 유전자를 분류하는 딥러닝 모델을 제시한다. 또한 The Cancer Genome Atlas(TCGA) 데이터를 통해서 모델 학습 후 기존 통계 및 머신러닝 기반 기법과 비교하여 성능이 개선되었음을 확인하였다.

• Journal of Information Science Theory and Practice
• /
• v.2 no.1
• /
• pp.6-21
• /
• 2014

This paper proposes a novel knowledge extraction system, TAKES (Two-step Approach for Knowledge Extraction System), which integrates advanced techniques from Information Retrieval (IR), Information Extraction (IE), and Natural Language Processing (NLP). In particular, TAKES adopts a novel keyphrase extraction-based query expansion technique to collect promising documents. It also uses a Conditional Random Field-based machine learning technique to extract important biological entities and relations. TAKES is applied to biological knowledge extraction, particularly retrieving promising documents that contain Protein-Protein Interaction (PPI) and extracting PPI pairs. TAKES consists of two major components: DocSpotter, which is used to query and retrieve promising documents for extraction, and a Conditional Random Field (CRF)-based entity extraction component known as FCRF. The present paper investigated research problems addressing the issues with a knowledge extraction system and conducted a series of experiments to test our hypotheses. The findings from the experiments are as follows: First, the author verified, using three different test collections to measure the performance of our query expansion technique, that DocSpotter is robust and highly accurate when compared to Okapi BM25 and SLIPPER. Second, the author verified that our relation extraction algorithm, FCRF, is highly accurate in terms of F-Measure compared to four other competitive extraction algorithms: Support Vector Machine, Maximum Entropy, Single POS HMM, and Rapier.

• IEIE Transactions on Smart Processing and Computing
• /
• v.5 no.6
• /
• pp.395-402
• /
• 2016

In these days, the wearable devices have been developed for measuring biological data effectively. However, wearable devices have tissue allege and noise problem. Also, it is impossible for a remote center to identify the person whose data are measured by wearable devices, which could trigger a communication problem over treatment. To solve these problems, biometric measurement based on a non-contact method, such as face image sequencing is necessary. This makes it possible to measure biometric data without any operation and side effects. This system can monitor the biological signals of people in real time without allege and noise and simultaneously identify them. In this paper, we propose an authentication process while measuring biometric data, through a non-contact method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.3
• /
• pp.771-791
• /
• 2022

Protein-protein interaction (PPI) extraction from original text is important for revealing the molecular mechanism of biological processes. With the rapid growth of biomedical literature, manually extracting PPI has become more time-consuming and laborious. Therefore, the automatic PPI extraction from the raw literature through natural language processing technology has attracted the attention of the majority of researchers. We propose a PPI extraction model based on the large pre-trained language model and adversarial training. It enhances the learning of semantic and syntactic features using BioBERT pre-trained weights, which are built on large-scale domain corpora, and adversarial perturbations are applied to the embedding layer to improve the robustness of the model. Experimental results showed that the proposed model achieved the highest F1 scores (83.93% and 90.31%) on two corpora with large sample sizes, namely, AIMed and BioInfer, respectively, compared with the previous method. It also achieved comparable performance on three corpora with small sample sizes, namely, HPRD50, IEPA, and LLL.

• The Journal of Fisheries Business Administration
• /
• v.17 no.1
• /
• pp.41-76
• /
• 1986

The northeastern Pacific and eastern Bering Sea supports large and divers finfish and shellfish compelxes. The North Pacific fishing grounds are largest fish production total 30.89 percent of the world in 1983. Recent action by the around countries extending their fisheies jurisdiction to 200 mile has provided a new environment. The new approch to fisheries management has resulted in an arormous expansion in tile demand for scientific information. These are no longer limited to biological concerns. Emphasis has been focused recently on the need to understand the economic, social, politics and philosophy characteristics of a fishing and how they will be affected by management decisions. The horizontal management and portfolio management for fisheries has increased interest in complex biological models and the coupling of these models with economic, politics and phyisophy components. Successful completion of this task will require and expanded understanding of oceanographic, biological, economic, social, politics and philosophic process associated with fisheries. Particular attention should be devoted to acquiring information an data processing for Korean trawl fishery, stock assessment in the areas. The need for international collaboration in management must be stressed. Some northeast Pacific and eastern Bering Sea fish stock migrate over long distances and political boundaries. Further, nearly all the fish stocks are harvested by several nations. The features require cooperation and coordination of research activities. The horizontal management will be made a way these activities for the around countries in the sea. Strongly, tile proposed again, Pacific International Council for Exploration of the Seas, PICES will interest between all users of the area's fisheries.