• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.3
• /
• pp.653-669
• /
• 2024

Dynamic scene deblurring is a complex computer vision problem owing to its difficulty to model mathematically. In this paper, we present a novel approach for image deblurring with the help of the sharp reference image, which utilizes the reference image for high-quality and high-frequency detail results. To better utilize the clear reference image, we develop an encoder-decoder network and two novel modules are designed to guide the network for better image restoration. The proposed Reference Extraction and Aggregation Module can effectively establish the correspondence between blurry image and reference image and explore the most relevant features for better blur removal and the proposed Spatial Feature Fusion Module enables the encoder to perceive blur information at different spatial scales. In the final, the multi-scale feature maps from the encoder and cascaded Reference Extraction and Aggregation Modules are integrated into the decoder for a global fusion and representation. Extensive quantitative and qualitative experimental results from the different benchmarks show the effectiveness of our proposed method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.3
• /
• pp.600-611
• /
• 1996

Recently, neural network-based speech recognition has been studied to utilize the adaptivity and learnability of neural network models. However, conventional neural network models have difficulty in the co-articulation processing and the boundary detection of similar phonmes of the Korean speech. Also, in case of using one phonotopic map, learning speed may dramatically increase and inaccuracies may be caused because homogeneous learning and recognition method should be applied for heterogenous data. Hence, in this paper, a neural net typewriter has been designed using a hierarchical self-organizing classifier(HSOC), and related algorithms are presented. This HSOC, during its learing stage, distributed phoneme data on hierarchically structured multiple phonotopic maps, using Kohonen's self-organizing feature maps(SOFM). Presented and experimented in this paper were the algorithms for deciding the number of maps, map sizes, the selection of phonemes and their placement per map, an approapriate learning and preprocessing method per map. If maps are divided according to a priorlinguistic knowledge, we would have difficulty in acquiring linguistic knowledge and how to alpply it(e.g., processing extended phonemes). Contrarily, our HSOC has an advantage that multiple phonotopic maps suitable for given input data are self-organizable. The resulting three korean phonotopic maps are optimally labelled and have their own optimal preprocessing schemes, and also confirm to the conventional linguistic knowledge.

• Ocean Systems Engineering
• /
• v.6 no.3
• /
• pp.217-231
• /
• 2016

This paper presents a new approach for underwater image analysis using the bi-dimensional empirical mode decomposition (BEMD) technique and the phase congruency information. The BEMD algorithm, fully unsupervised, it is mainly applied to texture extraction and image filtering, which are widely recognized as a difficult and challenging machine vision problem. The phase information is the very stability feature of image. Recent developments in analysis methods on the phase congruency information have received large attention by the image researchers. In this paper, the proposed method is called the EP model that inherits the advantages of the first two algorithms, so this model is suitable for processing underwater image. Moreover, the receiver operating characteristic (ROC) curve is presented in this paper to solve the problem that the threshold is greatly affected by personal experience when underwater image edge detection is performed using the EP model. The EP images are computed using combinations of the Canny detector parameters, and the binaryzation image results are generated accordingly. The ideal EP edge feature extractive maps are estimated using correspondence threshold which is optimized by ROC analysis. The experimental results show that the proposed algorithm is able to avoid the operation error caused by manual setting of the detection threshold, and to adaptively set the image feature detection threshold. The proposed method has been proved to be accuracy and effectiveness by the underwater image processing examples.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.4
• /
• pp.451-455
• /
• 2009

This paper provides a comparison of global path planning method in single string by using pulled and pushed SOFM (Self-Organizing Feature Map) which is a method among a number of neural network. The self-organizing feature map uses a randomized small valued initial-weight-vectors, selects the neuron whose weight vector best matches input as the winning neuron, and trains the weight vectors such that neurons within the activity bubble are move toward the input vector. On the other hand, the modified SOFM method in this research uses a predetermined initial weight vectors of the one dimensional string, gives the systematic input vector whose position best matches obstacles, and trains the weight vectors such that neurons within the activity bubble are move toward or reverse the input vector, by rising a pulled- or a pushed-SOFM. According to simulation results one can conclude that the modified neural networks in single string are useful tool for the global path planning problem of a mobile robot. In comparison of the number of iteration for converging to the solution the pushed-SOFM is more useful than the pulled-SOFM in global path planning for mobile robot.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.8
• /
• pp.1843-1859
• /
• 2013

The continuous increase of the spatial resolution of remote sensing images brings great challenge to image analysis and processing. Traditional prior knowledge-based region detection and target recognition algorithms for processing high resolution remote sensing images generally employ a global searching solution, which results in prohibitive computational complexity. In this paper, a more efficient region of interest (ROI) detection algorithm based on visual attention and threshold segmentation (VA-TS) is proposed, wherein a visual attention mechanism is used to eliminate image segmentation and feature detection to the entire image. The input image is subsampled to decrease the amount of data and the discrete moment transform (DMT) feature is extracted to provide a finer description of the edges. The feature maps are combined with weights according to the amount of the "strong points" and the "salient points". A threshold segmentation strategy is employed to obtain more accurate region of interest shape information with the very low computational complexity. Experimental statistics have shown that the proposed algorithm is computational efficient and provide more visually accurate detection results. The calculation time is only about 0.7% of the traditional Itti's model.

• Bulletin of the Korean Space Science Society
• /
• 2009.10a
• /
• pp.29.4-30
• /
• 2009

Extra-tidal feature of the globular clusters such as tidal tails and halos can be a crucial evidence of the merging scenario of the Galaxy formation in the dynamical point of view. To search for such an extra-tidal feature of globular clusters located in the Galactic bulge(RGC<3kpc), we obtained wide-field near-infrared JHKs images of 6 metal-poor ([Fe/H]<-1.0) clusters and 3 metal-rich ([Fe/H]>-1.0) clusters. Observations were carried out using IRSF 1.4m telescope and SIRIUS near-infrared camera, during 2006~2007. The obtained images have a total maximum field-of-view of ~ $21'\times 21'$. To select clusters' member stars and minimize the field star contaminations, we applied CMD masking algorithm. Smoothed surface density contour maps with selected stars for each cluster show overdensity features around the tidal radius and beyond. Also, radial surface density profiles within the tidal radius of the clusters show an overdensity feature as a change of slope of the radial profile. The results add further observational constraints of the formation of the Galactic bulge.

• The KIPS Transactions:PartB
• /
• v.16B no.2
• /
• pp.109-116
• /
• 2009

In this thesis, an object tracking method based on the active feature model and the optical flow in stereo images is proposed. We acquired the translation information of object of interest and the features of object by utilizing the geometric information and depth of stereo images. Tracking performance is improved for the occlude object with this information by predicting the movement information of features of the occlude object. Rigid and non-rigid objects are experimented. From the result of experiment, the OOI can be real-time tracked from complicate back ground. Besides, we got the improved result of object tracking in any occlusion state, no matter what it is rigid or non-rigid object.

• Journal of the Korea Society of Computer and Information
• /
• v.12 no.4
• /
• pp.229-237
• /
• 2007

To solve of complex problems for access and management data stored in the very large spatial database system that need to constructed metadata for the physical and logical elements concerned with spatial and non-spatial data sets. Also, in the underground water management system which managed with spatial and non-spatial elements, need to spatial features metadata system based on feature-based. We, in this paper, proposed metadata model for the feature-based underground water management system using underground water meta-information inputted on the DXF formatted tile-based geological maps. Additional, we modeled metadata level of feature and data set and presented standard specification of underground water metadata.

• 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.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.3
• /
• pp.1434-1449
• /
• 2019

Video object segmentation is a significant task in computer vision, but its performance is not very satisfactory. A method of video object segmentation using weakly temporal information is presented in this paper. Motivated by the phenomenon in reality that the motion of the object is a continuous and smooth process and the appearance of the object does not change much between adjacent frames in the video sequences, we use a feed-forward architecture with motion estimation to predict the mask of the current frame. We extend an additional mask channel for the previous frame segmentation result. The mask of the previous frame is treated as the input of the expanded channel after processing, and then we extract the temporal feature of the object and fuse it with other feature maps to generate the final mask. In addition, we introduce multi-mask guidance to improve the stability of the model. Moreover, we enhance segmentation performance by further training with the masks already obtained. Experiments show that our method achieves competitive results on DAVIS-2016 on single object segmentation compared to some state-of-the-art algorithms.