• 한국컴퓨터그래픽스학회논문지
• /
• 제30권3호
• /
• pp.51-59
• /
• 2024

시선 벡터 정답값을 갖는 대규모 데이터의 수집은 시선 추적 분야에서 많은 비용을 필요로 한다. 본 논문에서는 원본 사진의 시선을 수정하는 데이터 증강 기법을 사용하여 제한된 개수의 시선 정답값이 주어진 상황에서 시선 추적 모델의 정확도를 향상시키는 방법을 제안한다. 시선 구간 다중 클래스 분류를 보조 작업으로 학습하고, 디퓨전 오토인코더의 잠재 변수를 조정하여 원본 사진의 시선을 편집한 사진을 생성한다. 기존의 얼굴 속성 편집과 달리, 우리는 이진 속성이 아닌 시선 벡터의 피치와 요를 지정한 범주 내로 변경하며, 편집된 사진을 시선 추적 모델의 증강된 학습 데이터로 활용한다. 시선 정답값이 5만 개 이하일 때 준지도 학습에서의 시선 추적 모델의 정확도 향상은 제안한 데이터 증강 기법의 효과를 입증한다.

• 식물조직배양학회지
• /
• 제22권2호
• /
• pp.95-99
• /
• 1995

본 연구는 이미 확립되어 있는 고려인삼의 체세포배발생을 통한 식물체 재분화와 Agrobacterium을 매개로 한 형질전환 시스템을 이용하여 항곰팡이성 인삼을 개발하고자 염기성인 강낭콩 키틴가수분해효소 유전자를 인삼으로 도입하였다. CaMV 35S promoter-강낭콩 키틴가수분해효소 유전자와 선발표지로서의 neomycin phosphotransferase II (NPT II) 유전자를 가진 pChi/748 binary 벡터를 pGA748로부터 제조하여 이를 도입한 A. tumefacience LBA4404와 인삼 접합배의 자엽절편을 1 mg/L 24-D, 0.1 mg/L kinetin이 첨가된 MS 액체배지에서 48시간 동안 공동배양한 후 동일배지에 100 mg/L kanamycine 500 mg/L carbenicillin을 첨가한 고체 배지에 옮겨 배양하였다. 배양 한달 후부터 절편의 절단면 부근으로부터 캘러스가 유도되기 시작하였으며 이어서 수많은 체세포배가 형성되었다. 이들 체세포배를 BA와 GA3가 각각 1 mg/L 첨가된 배지로 옮겨서 5주 경과되었을 때 식물체로 전환되었다. 재분화된 개체 중 선발된 8개의 식물체로부터 PCR과 이 산물의 Southern분석 결과 6개의 재분화 개체에서 강낭콩 키틴가수분해효소 유전자가 도입되었음을 확인하였다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.1889-1891
• /
• 2004

The differential white blood cell count plays an important role in the diagnosis of different diseases. It is a tedious task to count these classes of cell manually. An automatic counter using computer vision helps to perform this medical test rapidly and accurately. Most commercial-available automatic white blood cell analysis composed mainly 3 steps including segmentation, feature extraction and classification. In this paper we concentrate on the first step in automatic white-blood-cell analysis by proposing a segmentation scheme that utilizes a benefit of active contour. Specifically, the binary image is obtained by thresolding of the input blood smear image. The initial shape of active is then placed roughly inside the white blood cell and allowed to grow to fit the shape of individual white blood cell. The white blood cell is then separated using the extracted contour. The force that drives the active contour is the combination of gradient vector flow force and balloon force. Our purposed technique can handle very promising to separate the remaining red blood cells.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.312-316
• /
• 2004

The paper presents a technique to identify the boundary of the optic disc in infant retinal digital images using an approach based on active contours (snakes). The technique can be used to be develop a automate system in order to help the ophthalmologist's diagnosis the retinopathy of prematurity (ROP) disease which may occurred on preterm infant,. The optic disc detection is one of the fundamental step which could help to create an automate diagnose system for the doctors we use a new kind of active contour (snake) method has been developed by Chenyang et. al. [1], based on a new type of external force field, called gradient vector flow, or GVF. GVF is computed as a diffusion of the gradient vectors of a gray-level or binary edge map derived from the image. The testing results on a set of infant retinal ROP images verify the effectiveness of the proposed methods. We show that GVF has a large capture range and it's able to move snakes into boundary concavities of optic disc and finally the optic disk boundary was determined.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권9호
• /
• pp.3762-3781
• /
• 2020

Infrared pedestrian target recognition is a vital problem of significant interest in computer vision. In this work, a novel infrared pedestrian target recognition method that uses heterogeneous features with multi-kernel transfer learning is proposed. Firstly, to exploit the characteristics of infrared pedestrian targets fully, a novel multi-scale monogenic filtering-based completed local binary pattern descriptor, referred to as MSMF-CLBP, is designed to extract the texture information, and then an improved histogram of oriented gradient-fisher vector descriptor, referred to as HOG-FV, is proposed to extract the shape information. Second, to enrich the semantic content of feature expression, these two heterogeneous features are integrated to get more complete representation for infrared pedestrian targets. Third, to overcome the defects, such as poor generalization, scarcity of tagged infrared samples, distributional and semantic deviations between the training and testing samples, of the state-of-the-art classifiers, an effective multi-kernel transfer learning classifier called MK-TrAdaBoost is designed. Experimental results show that the proposed method outperforms many state-of-the-art recognition approaches for infrared pedestrian targets.

• 한국멀티미디어학회논문지
• /
• 제20권8호
• /
• pp.1395-1405
• /
• 2017

In this paper, we analyses heart beat data to identify subjects stress state (binary) using heart rate variability (HRV) features extracted from heart beat data of the subjects and implement supervised machine learning techniques to create the mental stress classifier. There are four steps need to be done: data acquisition, data processing (HRV analysis), features selection, and machine learning, before doing performance measurement. There are 56 features generated from the HRV Analysis module with several of them are selected (using own algorithm) after computing the Pearson Correlation Matrix (p-values). The results of the list of selected features compared with all features data are compared by its model error after training using several machine learning techniques: support vector machine, decision tree, and discriminant analysis. SVM model and decision tree model with using selected features shows close results compared to using all recording by only 1% difference. Meanwhile, the discriminant analysis differs about 5%. All the machine learning method used in this works have 90% maximum average accuracy.

• ETRI Journal
• /
• 제40권3호
• /
• pp.396-409
• /
• 2018

Elliptic curve cryptography (ECC) can achieve relatively good security with a smaller key length, making it suitable for Internet of Things (IoT) devices. DNA-based encryption has also been proven to have good security. To develop a more secure and stable cryptography technique, we propose a new hybrid DNA-encoded ECC scheme that provides multilevel security. The DNA sequence is selected, and using a sorting algorithm, a unique set of nucleotide groups is assigned. These are directly converted to binary sequence and then encrypted using the ECC; thus giving double-fold security. Using several examples, this paper shows how this complete method can be realized on IoT devices. To verify the performance, we implement the complete system on the embedded platform of a Raspberry Pi 3 board, and utilize an active sensor data input to calculate the time and energy required for different data vector sizes. Connectivity and resilience analysis prove that DNA-mapped ECC can provide better security compared to ECC alone. The proposed method shows good potential for upcoming IoT technologies that require a smaller but effective security system.

• Smart Structures and Systems
• /
• 제18권6호
• /
• pp.1233-1250
• /
• 2016

Cracks in plate-like structures are some of the main reasons for destruction of the entire structure. In this study, a novel two-stage methodology is proposed for damage detection of flexural plates using an optimized artificial neural network. In the first stage, location of damages in plates is investigated using curvature-moment and curvature-moment derivative concepts. After detecting the damaged areas, the equations for damage severity detection are solved via Bat Algorithm (BA). In the second stage, in order to efficiently reduce the computational cost of model updating during the optimization process of damage severity detection, multiple damage location assurance criterion index based on the frequency change vector of structures are evaluated using properly trained cascade feed-forward neural network (CFNN) as a surrogate model. In order to achieve the most generalized neural network as a surrogate model, its structure is optimized using binary version of BA. To validate this proposed solution method, two examples are presented. The results indicate that after determining the damage location based on curvature-moment derivative concept, the proposed solution method for damage severity detection leads to significant reduction of computational time compared with direct finite element method. Furthermore, integrating BA with the efficient approximation mechanism of finite element model, maintains the acceptable accuracy of damage severity detection.

• KSBB Journal
• /
• 제16권5호
• /
• pp.480-485
• /
• 2001

국화과에 속하는 다년생 초본식물의 민들레가 Agrobacteria 에 대한 숙주로서의 가능성을 조사하고 여러 가지 유용한 유전자를 민들레로 도입시키기 위해, 민들레잎 절편을 pBI121으로 형질전환된 Agrobacterium tumefaciens LBA4404와 10분동안 공동배양하여 형질전환시킨 후, 1$\mu$M IAA, 1$\mu$M BA. 50 $\mu$g/ML Km과 100$\mu$g/ML Cb이 함유된 MS 배지에서 약 2주후에 multiple shoot를 유가시켰다. 유기된 shoot로 부터 유식식물체를 얻었으며, 형질전환을 확인하기 위해 GUS활성을 측정한 결과 양성 반응을 보였다.

• Journal of Information Processing Systems
• /
• 제9권1호
• /
• pp.173-188
• /
• 2013

In Facial Expression Recognition Systems (FERS), only particular regions of the face are utilized for discrimination. The areas of the eyes, eyebrows, nose, and mouth are the most important features in any FERS. Applying facial features descriptors such as the local binary pattern (LBP) on such areas results in an effective and efficient FERS. In this paper, we propose an automatic facial expression recognition system. Unlike other systems, it detects and extracts the informative and discriminant regions of the face (i.e., eyes, nose, and mouth areas) using Haar-feature based cascade classifiers and these region-based features are stored into separate image files as a preprocessing step. Then, LBP is applied to these image files for facial texture representation and a feature-vector per subject is obtained by concatenating the resulting LBP histograms of the decomposed region-based features. The one-vs.-rest SVM, which is a popular multi-classification method, is employed with the Radial Basis Function (RBF) for facial expression classification. Experimental results show that this approach yields good performance for both frontal and near-frontal facial images in terms of accuracy and time complexity. Cohn-Kanade and JAFFE, which are benchmark facial expression datasets, are used to evaluate this approach.