• 한국정보통신학회논문지
• /
• 제6권6호
• /
• pp.819-826
• /
• 2002

셀룰러 뉴럴 네트워크는 국부적 연결특성을 가지고 있어 실시간 이미지처리에 적합한 뉴럴 네크워크이다. 또한 국부적 연결특징은 VLSI구현에 적합하다. 그의 응용분야는 패턴인식과 숫자인식 및 영상처리에 응용되고 있다. 본 논문에서, CNN은 전처리 단계로서 숫자의 특징점 추출에 이용된다. CNN을 이용한 그림자검출은 4내지 6방향으로 검출하여 숫자의 특징점을 방향별로 추출한다. 분류단계에서 이러한 형상자료는 다층BP뉴럴 네트워크의 입력벡터에 적합하도록 압축되어 입력된다. 실험결과 CNN을 통한 숫자인식은 굴림체의 경우96%이상의 인식율을 보여 만족할 만한 결과를 얻었다.

• 한국지능시스템학회논문지
• /
• 제13권3호
• /
• pp.371-376
• /
• 2003

본 논문에서는 효과적인 정보를 표현하는 Independent Component Analysis(ICA)-factorial 표현방법을 이용하여 얼굴감정 인식을 수행한다. 얼굴감정인식은 두 단계인 특징추출 과정과 인식과정에 의해 이루어진다. 먼저 특징추출방법은 주성분 분석(Principal Component Analysis)을 이용하여 얼굴영상의 고차원 공간을 저차원 특징공간으로 변환한 후 ICA-factorial 표현방법을 통해 좀 더 효과적으로 특징벡터를 추출한다. 인식단계는 최소거리 분류방법인 유클리디안 거리에 근거한 K-Nearest Neighbor 알고리즘으로 얼굴감정을 인식한다. 6개의 기본감정(기쁨, 슬픔, 화남, 놀람, 공포, 혐오)에 대해 얼굴 감정 데이터베이스를 구축하고 실험해본 결과 기존의 방법보다 좋은 인식 성능을 얻었다.

• International Journal of Computer Science & Network Security
• /
• 제22권10호
• /
• pp.1-10
• /
• 2022

Nowadays software defect prediction (SDP) is most active research going on in software engineering. Early detection of defects lowers the cost of the software and also improves reliability. Machine learning techniques are widely used to create SDP models based on programming measures. The majority of defect prediction models in the literature have problems with class imbalance and high dimensionality. In this paper, we proposed Centroid and Nearest Neighbor based Class Imbalance Reduction (CNNCIR) technique that considers dataset distribution characteristics to generate symmetry between defective and non-defective records in imbalanced datasets. The proposed approach is compared with SMOTE (Synthetic Minority Oversampling Technique). The high-dimensionality problem is addressed using Ant Colony Optimization (ACO) technique by choosing relevant features. We used nine different classifiers to analyze six open-source software defect datasets from the PROMISE repository and seven performance measures are used to evaluate them. The results of the proposed CNNCIR method with ACO based feature selection reveals that it outperforms SMOTE in the majority of cases.

• 한국컴퓨터정보학회논문지
• /
• 제27권5호
• /
• pp.77-84
• /
• 2022

본 논문에서는 새로운 비지도 특징 선별 기법을 제안한다. 기존 비지도 방식의 특징 선별 기법들은 특징을 선별하기 위해 가상의 레이블 데이터를 정하고 주어진 데이터를 이 레이블 데이터에 사영하는 회귀 분석 방식으로 특징을 선별하였다. 하지만 가상의 레이블은 데이터로부터 생성되기 때문에 사영된 공간이 비슷하게 형성될 수 있다. 따라서 기존의 방법들에서는 제한된 공간에서만 특징이 선택될 수 있었다. 이를 해소하기 위해 본 논문에서는 직교 사영과 저랭크 근사를 이용하여 특징을 선별한다. 이 문제를 해소하기 위해 가상의 레이블을 직교 사영하고 이 공간에 데이터를 사영할 수 있도록 한다. 이를 통해 더 주요한 특징 선별을 기대할 수 있다. 그리고 사영을 위한 변환 행렬에 저랭크 제한을 두어 더 효과적으로 저차원 공간의 특징을 선별할 수 있도록 한다. 이 목표를 달성하기 위해 본 논문에서는 비용 함수를 설계하고 효율적인 최적화 방법을 제안한다. 여섯 개의 데이터에 대한 실험 결과는 제안된 방법이 대부분의 경우 기존의 비지도 특징 선별 기법보다 좋은 성능을 보여주었다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권10호
• /
• pp.2768-2787
• /
• 2023

As an important research direction of the application of computer science in the medical field, the automatic generation technology of radiology report has attracted wide attention in the academic community. Because the proportion of normal regions in radiology images is much larger than that of abnormal regions, words describing diseases are often masked by other words, resulting in significant feature loss during the calculation process, which affects the quality of generated reports. In addition, the huge difference between visual features and semantic features causes traditional multi-modal fusion method to fail to generate long narrative structures consisting of multiple sentences, which are required for medical reports. To address these challenges, we propose a multi-feature optimization Transformer (MFOT) for generating radiology reports. In detail, a multi-dimensional mapping attention (MDMA) module is designed to encode the visual grid features from different dimensions to reduce the loss of primary features in the encoding process; a feature pre-fusion (FP) module is constructed to enhance the interaction ability between multi-modal features, so as to generate a reasonably structured radiology report; a detail enhanced attention (DEA) module is proposed to enhance the extraction and utilization of key features and reduce the loss of key features. In conclusion, we evaluate the performance of our proposed model against prevailing mainstream models by utilizing widely-recognized radiology report datasets, namely IU X-Ray and MIMIC-CXR. The experimental outcomes demonstrate that our model achieves SOTA performance on both datasets, compared with the base model, the average improvement of six key indicators is 19.9% and 18.0% respectively. These findings substantiate the efficacy of our model in the domain of automated radiology report generation.

• The Journal of the Acoustical Society of Korea
• /
• 제17권3E호
• /
• pp.12-21
• /
• 1998

In this paper, we describe the one channel five-way, V/U/M/N/S (Voice/Unvoice/Nasal/Silent), classification algorithm for automatically classifying speech. The decision making process is viewed as a pattern viewed as a pattern recognition problem. Two aspects of the algorithm are developed: feature selection and classifier type. The feature selection procedure is studied for identifying a set of features to make V/U/M/N/S classification. The classifiers used are a vector quantization (VQ), a neural network(NN), and a decision tree method. Actual five sentences spoken by six speakers, three male and three female, are tested with proposed classifiers. From a set of measurement tests, the proposed classifiers show fairly good accuracy for V/U/M/N/S decision.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제15권4호
• /
• pp.277-282
• /
• 2015

Braincomputer interface (BCI) technology is making advances in the field of humancomputer interaction (HCI). To improve the BCI technology, we study the changes in the electroencephalogram (EEG) signals for six levels of grip strength: 10%, 20%, 40%, 50%, 70%, and 80% of the maximum voluntary contraction (MVC). The measured EEG data are categorized into three classes: Weak, Medium, and Strong. Features are then extracted using power spectrum analysis and multiclass-common spatial pattern (multiclass-CSP). Feature datasets are classified using a support vector machine (SVM). The accuracy rate is higher for the Strong class than the other classes.

• 음성과학
• /
• 제12권3호
• /
• pp.201-209
• /
• 2005

This study was investigated developmental temporal features in Putonghua-speaking children. The total of 212 children between the ages 2;6 and 6;5 participated in Shanghai. Speech materials were constructed according to aspiration feature in stop sounds of Putonghua. Six words were selected in this study. A voice onset time was measured. Non-parametric procedures were employed for all the analyses. The VOT value across bilabial, alveolar, and velar stops was significantly differed between aspirated and unaspirated stops for each age group. Effect of age is. significant for unaspirated stops. It is clear that each of Putonghua stops showed decreasing mean and standard deviation. The overshoot phenomenon of VOT was apparent from the age of 2;6-2;11 to 4;6-4;11. There was high variability in the production of lag time for aspirated stops.

• International Journal of CAD/CAM
• /
• 제6권1호
• /
• pp.117-123
• /
• 2006

Recently with the development of 3D modeling and digitizing tools, more and more models have been created, which leads to the necessity of the technique of 3D mode retrieval system. In this paper we investigate a new method for 3D model retrieval based on orthogonal projections. We assume that 3D models are composed of trigonal meshes. Algorithms process first by a normalization step in which the 3D models are transformed into the canonical coordinates. Then each model is orthogonally projected onto six surfaces of the projected cube which contains it. A following step is feature extraction of the projected images which is done by Moment Invariants and Polar Radius Fourier Transform. The feature vector of each 3D model is composed of the features extracted from projected images with different weights. Our System validates that this means can distinguish 3D models effectively. Experiments show that our method performs quit well.

• 전자공학회논문지B
• /
• 제31B권9호
• /
• pp.142-149
• /
• 1994

Hangul is expressed by the basic elements, twenty-four characters. Because these characters are composed of a circle and lines, Hough transform(HT), which has a powerful performance on the noise in extracting lines, is introduced. Many difficulties often occur when the original HT is used to extract strokes and it's direction, position and length from handwritten Hangul characters. Original HT has eight direction selected as samples in the transformed image should be calculated for these eight directions. In this paper, the generalized sampling rule is suggested. According to the rule, those directions which are possible to a line are the only thing to be calculated. The experoment result turned out to be higher than the method that Chen suggested in sampling rate. Anogher experiment result is done on the 1800 handwritten Hangul characters that 10 persons wrote. By feature extracting the oritinal HT and sampling HT. And as a result of six type classification, the suggested method came out higher than original HT.