• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.808-810
• /
• 2004

본 논문에서는 Contrast map과 Salient point를 이용하여 영상에서 중요한 객체를 자동으로 추출하는 방법을 제안한다. 우선 인간의 시각 체계와 유사한 밝기(luminance), 색상(color) 그리고 방향성(orientation) 3가지의 특징정보를 이용하여 각각의 특징정보로부터 feature map을 생성하고 이 3가지의 feature map을 선형 결합하여 contrast map을 생성한다. 이렇게 생성된 하나의 contrast map을 이용하여 대략적인 Attention Window (AW)의 위치를 결정한다. 다음으로, 영상으로부터 웨이블릿 변환을 적용하여 salient point를 찾고, salient point의 분포와 contrast map의 중요도에 따라 AW의 크기를 실제 중요 객체의 크기와 가장 유사하도록 축소시킨다. 이렇게 선택되고 축소된 AW안에서 실제 중요 객체를 추출하기 위해 AW 내부에 존재하는 영상에 대해서만 영상 분할을 하고 불필요한 영역을 제거하여 자동으로 중요객체를 추출하도록 한다.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.5
• /
• pp.441-446
• /
• 2020

In this paper, we propose the α-Feature Map Scaling (α-FMS) method which extends the FMS method that was designed to enhance the discriminative power of feature maps of deep neural networks in Speaker Verification (SV) systems. The FMS derives a scale vector from a feature map and then adds or multiplies them to the features, or sequentially apply both operations. However, the FMS method not only uses an identical scale vector for both addition and multiplication, but also has a limitation that it can only add a value between zero and one in case of addition. In this study, to overcome these limitations, we propose α-FMS to add a trainable parameter α to the feature map element-wise, and then multiply a scale vector. We compare the performance of the two methods: the one where α is a scalar, and the other where it is a vector. Both α-FMS methods are applied after each residual block of the deep neural network. The proposed system using the α-FMS methods are trained using the RawNet2 and tested using the VoxCeleb1 evaluation set. The result demonstrates an equal error rate of 2.47 % and 2.31 % for the two α-FMS methods respectively.

• The KIPS Transactions:PartB
• /
• v.8B no.5
• /
• pp.515-522
• /
• 2001

Feature-based similarity retrieval become an important research issue in image database systems. The features of image data are useful to discrimination of images. In this paper, we propose the highspeed k-Nearest Neighbor search algorithm based on Self-Organizing Maps. Self-Organizing Map(SOM) provides a mapping from high dimensional feature vectors onto a two-dimensional space. A topological feature map preserves the mutual relations (similarity) in feature spaces of input data, and clusters mutually similar feature vectors in a neighboring nodes. Each node of the topological feature map holds a node vector and similar images that is closest to each node vector. We implemented about k-NN search for similar image classification as to (1) access to topological feature map, and (2) apply to pruning strategy of high speed search. We experiment on the performance of our algorithm using color feature vectors extracted from images. Promising results have been obtained in experiments.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2475-2478
• /
• 2002

This paper consider's the problem of exploring an unfamiliar environment in search of recognizable objects of visual landmarks. In order to extract and recognize them automatically, a feature map is constructed which records the set of features continually during a learning phase. The map contains photometric geometric, and metric information of each feature. Meanwhile, the localization algorithm can determine the position of the robot by extracting features and matching in the map. These procedures are implemented and tested using an AMR, and preliminary results are presented in this paper.

• Journal of Korea Multimedia Society
• /
• v.18 no.4
• /
• pp.460-472
• /
• 2015

Most of previous visual attention system finds attention regions based on saliency map which is combined by multiple extracted features. The differences of these systems are in the methods of feature extraction and combination. This paper presents a new system which has an improvement in feature extraction method of color and motion, and in weight decision method of spatial and temporal features. Our system dynamically extracts one color which has the strongest response among two opponent colors, and detects the moving objects not moving pixels. As a combination method of spatial and temporal feature, the proposed system sets the weight dynamically by each features' relative activities. Comparative results show that our suggested feature extraction and integration method improved the detection rate of attention region.

• Journal of the Korean Society of Industry Convergence
• /
• v.24 no.1
• /
• pp.1-7
• /
• 2021

In this paper, we presented the experimental basis for the theoretical background and robustness of the Convolutional Neural Network for object recognition based on artificial intelligence. An experimental result was performed to visualize the weighting filters and feature maps for each layer to determine what characteristics CNN is automatically generating. experimental results were presented on the trend of learning error and identification error rate by checking the relevance of the weight filter and feature map for learning error and identification error. The weighting filter and characteristic map are presented as experimental results. The automatically generated characteristic quantities presented the results of error rates for moving and rotating robustness to geometric changes.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2003.10a
• /
• pp.281-286
• /
• 2003

Since National Geographic Information System was started, paper maps have been made with computer aided editing of digital map, instead of etching map-size negative film. Automated paper mapping system's necessity is growing more and more, because digital map has changed into Ver.2.0 which include attributes of feature. Therefore, in this study we try to analyze correlation of the digital map feature code and the 1/5,000 topographic map specifications which is necessary for paper mapping automatization using digital map Ver.2.0, and try to develop fundamental modules which will play a core role in automated paper mapping system.

• Journal of Korea Multimedia Society
• /
• v.19 no.9
• /
• pp.1647-1658
• /
• 2016

RGB+D database has been widely used in object recognition, object tracking, robot control, to name a few. While rapid advance of active depth sensing technologies allows for the widespread of indoor RGB+D databases, there are only few outdoor RGB+D databases largely due to an inherent limitation of active depth cameras. In this paper, we propose a novel method used to build outdoor RGB+D databases. Instead of using active depth cameras such as Kinect or LIDAR, we acquire a pair of stereo image using high-resolution stereo camera and then obtain a depth map by applying stereo matching algorithm. To deal with estimation errors that inevitably exist in the depth map obtained from stereo matching methods, we develop an approach that estimates confidence of depth maps based on unsupervised learning. Unlike existing confidence estimation approaches, we explicitly consider a spatial correlation that may exist in the confidence map. Specifically, we focus on refining confidence feature with the assumption that the confidence feature and resultant confidence map are smoothly-varying in spatial domain and are highly correlated to each other. Experimental result shows that the proposed method outperforms existing confidence measure based approaches in various benchmark dataset.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.394-394
• /
• 2000

This Paper proposes a new nonlinear partial least square method that extends the linear PLS. Proposed nonlinear PLS uses self-organizing feature map as PLS outer relation and multilayer neural network as PLS inner regression method.

• Journal of the Korea Society of Computer and Information
• /
• v.8 no.2
• /
• pp.31-36
• /
• 2003

This paper proposes an acceleration method of PCA(Principal Component Analysis) based feature detection. The feature detection method makes decision whether the target feature is included in a given image, and if included, calculates the position and extent of the target feature. The position and scale of the target feature or face is not known previously, all the possible locations should be tested for various scales to detect the target. This is a search Problem in huge search space. This Paper proposes a fast face and feature detection method by reducing the search space using the multi-stage prediction map and contour Prediction map. A Proposed method compared to the existing whole search way, and it was able to reduce a computational complexity below 10％ by experiment.