• Journal of the Korea Society of Computer and Information
• /
• v.21 no.11
• /
• pp.49-57
• /
• 2016

In this paper, we propose the novel tracking method which uses the weighted discriminative correlation filter (DCF). We also propose the PSPR instead of conventional PSR as tracker performance evaluation method. The proposed tracking method uses multiple DCF to estimates the target position. In addition, our proposed method reflects more weights on the correlation response of the tracker which is expected to have more performance using PSPR. While existing multi-DCF-based tracker calculates the final correlation response by directly summing correlation responses from each tracker, the proposed method acquires the final correlation response by weighted combining of correlation responses from the selected trackers robust to given environment. Accordingly, the proposed method can provide high performance tracking in various and complex background compared to multi-DCF based tracker. Through a series of tracking experiments for various video data, the presented method showed better performance than a single feature-based tracker and also than a multi-DCF based tracker.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.34C no.1
• /
• pp.59-64
• /
• 1997

By using a selective learnable self-organizing feature map(SOFM) a more practical and generalized mehtod is proposed in which the effective nonlinear shape restoration is possible regardless of the existence of the distortion modelss. Nonlinear mapping relation is extracted from the distorted imate by using the proposed selective learning SOFGM which has the special property of effectively creating spatially organized internal representations and nonlinear relations of various input signals. For the exact extraction of the mapping relations between the distorted image and the original one, we define a disparity index as a proximal nmeasure of the present state to the final idealy trained state of the SOFM, and we used this index to adjust the training of the mapping relations form the weights of the SOFM. Simulations are conducted on various kinds of distorted images with or without distortion models, and the results show that the proposed method is very efficeint very efficient and practical in nonlinear shape restorations.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.41 no.5
• /
• pp.533-540
• /
• 1992

The Self Creating and Organizing (SCO) is a new architecture and one of the unsupervized learning algorithm for the artificial neural network. SCO begins with only one output node which has a sufficiently wide response range, and the response ranges of all the nodes decrease automatically whether adapting the weights of existing node or creating a new node. It is compared to the Kohonen's Self Organizing Feature Map (SOFM). The results show that SCONN has lots of advantages over other competitive learning architecture.

• Proceedings of the KIEE Conference
• /
• 2003.07d
• /
• pp.2516-2518
• /
• 2003

본 논문은 고차 뉴런의 문제점으로 지적되고 있는 뉴런이 방대하게 증가하는 문제를 해결하고자, 최적의 뉴런을 생성하고 생성되어진 고차 뉴런 중 일정 비율로 뉴런의 연결강도를 도태시켜 감에 따라 네트워크상에 나타나는 특성을 비교하였다. 본 논문은 고차 뉴런을 이용한 Kohonen의 자기 조직화 맵의 고차 뉴런부에 일정 비율로 연결강도를 도태한 후 인식률을 얻는 형태로 시뮬레이션을 하였다. 특히, 종래 형태의 고차 뉴런을 이용한 Kohonen 자기 조직화 맵의 알고리즘을 변형없이 사용하였으며 중복되는 뉴런을 최대한 억제하기 위해 2차 뉴런만을 생성한 네트워크 구조 위에 입력 데이터의 특징을 유지하고 고차 뉴런의 특징을 더욱 활성화하기 위해 일정한 양의 연결강도를 도태시킴으로써 출력면에서 국소집중 반응에 의한 정확한 인식률 향상 등을 조사하는 시뮬레이션을 하였다. 본 제안 모델의 특성을 살펴보기 위해 60개의 데이터로 이루어진 금속 소나 음데이터와 암석 소나 음 데이터를 이용하여 금속인지 암석인지를 판별하는 시뮬레이션을 하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10a
• /
• pp.130-132
• /
• 2004

본 연구에서는 FMM 신경망을 이용한 패턴 분류 문제에서 학습 패턴에 포함되는 특징의 발생 빈도와 특징 값의 분포를 고려하는 네트워크 구조와 학습 방법론을 소개한다. 이를 위하여 하이퍼박스 소속함수의 산출 과정에 세부특징에 대한 가중치 개념이 적용되는 새로운 활성화 특성을 제안한다. 또한 하이퍼박스의 특징 범위와 빈도 및 특징 값의 분포를 유지하고 새롭게 정의된 하이퍼박스 생성, 확장, 축소기법을 적용한다 이는 가중치 개념을 통하여 각 특징별 중요도를 서로 다른 값으로 반영할 수 있게 하며, 특징의 분포 정보가 고려되어 기존 FMM 모델에 비하여 노이즈에 의한 영향을 개선하여 학습 효과를 증진시킬 뿐만 아니라 하이퍼박스의 생성 및 확장 과정 중에 학습패턴의 순서에 상관없이 동일한 특성을 보일 수 있게 한다.

• Annual Conference on Human and Language Technology
• /
• 2010.10a
• /
• pp.63-67
• /
• 2010

본 논문에서는 자질(feature)의 가중치를 학습하여 이용하는 기계학습 기반 한국어 의존 파싱 기법을 소개한다. 이를 위하여 모든 가능한 의존관계에 대하여 각 의존관계마다 일정한 수의 자질을 생성한다. 자질마다 가중치에 의하여 그 중요도를 나타낸다. 자질의 가중치 값은 의존관계가 태깅된 구문구조 학습 말뭉치를 이용하여 학습한다. 이를 위해 본 논문에서는 간단한 가중치 기계학습 기법을 제시한다. 실험을 위한 언어 자원으로는 구구조부착 세종말뭉치를 변환하여 구한 의존관계 부착 말뭉치를 사용하였다. 실험 결과 약 86.5%의 정확률을 가지는 의존파싱이 가능함을 관찰하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.6
• /
• pp.2115-2127
• /
• 2021

Although significant progress has been made in synthesizing visually realistic face images by Generative Adversarial Networks (GANs), there still lacks effective approaches to provide fine-grained control over the generation process for semantic facial attribute editing. In this work, we propose a novel cross channel self-attention based generative adversarial network (CCA-GAN), which weights the importance of multiple channels of features and archives pixel-level feature alignment and conversion, to reduce the impact on irrelevant attributes while editing the target attributes. Evaluation results show that CCA-GAN outperforms state-of-the-art models on the CelebA dataset, reducing Fréchet Inception Distance (FID) and Kernel Inception Distance (KID) by 15~28% and 25~100%, respectively. Furthermore, visualization of generated samples confirms the effect of disentanglement of the proposed model.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.1
• /
• pp.166-179
• /
• 2021

Recently, with the development of technology, computer vision research based on the human visual system has been actively conducted. Saliency maps have been used to highlight areas that are visually interesting within the image, but they can suffer from low performance due to external factors, such as an indistinct background or light source. In this study, existing color, brightness, and contrast feature maps are subjected to multiple shape and orientation filters and then connected to a fully connected layer to determine pixel intensities within the image based on location-based weights. The proposed method demonstrates better performance in separating the background from the area of interest in terms of color and brightness in the presence of external elements and noise. Location-based weight normalization is also effective in removing pixels with high intensity that are outside of the image or in non-interest regions. Our proposed method also demonstrates that multi-filter normalization can be processed faster using parallel processing.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.3
• /
• pp.816-839
• /
• 2023

In a random vector functional link (RVFL) network, shortcomings such as local optimal stagnation and decreased convergence performance cause a reduction in the accuracy of illumination correction by only inputting the weights and biases of hidden neurons. In this study, we proposed an improved regularized random vector functional link (RRVFL) network algorithm with an optimized grey wolf optimizer (GWO). Herein, we first proposed the moth-flame optimization (MFO) algorithm to provide a set of excellent initial populations to improve the convergence rate of GWO. Thereafter, the MFO-GWO algorithm simultaneously optimized the input feature, input weight, hidden node and bias of RRVFL, thereby avoiding local optimal stagnation. Finally, the MFO-GWO-RRVFL algorithm was applied to ameliorate the performance of illumination correction of various test images. The experimental results revealed that the MFO-GWO-RRVFL algorithm was stable, compatible, and exhibited a fast convergence rate.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.3
• /
• pp.721-739
• /
• 2023

Rating prediction is an important issue in recommender systems, and its accuracy affects the experience of the user and the revenue of the company. Traditional recommender systems use Factorization Machinesfor rating predictions and each feature is selected with the same weight. Thus, there are problems with inaccurate ratings and limited data representation. This study proposes a deep recommendation model based on self-attention Factorization (SAFMR) to solve these problems. This model uses Convolutional Neural Networks to extract features from user and item reviews. The obtained features are fed into self-attention mechanism Factorization Machines, where the self-attention network automatically learns the dependencies of the features and distinguishes the weights of the different features, thereby reducing the prediction error. The model was experimentally evaluated using six classes of dataset. We compared MSE, NDCG and time for several real datasets. The experiment demonstrated that the SAFMR model achieved excellent rating prediction results and recommendation correlations, thereby verifying the effectiveness of the model.