• Journal of Information Processing Systems
• /
• 제19권4호
• /
• pp.427-438
• /
• 2023

Plenty of works have indicated that single image super-resolution (SISR) models relying on synthetic datasets are difficult to be applied to real scene text image super-resolution (STISR) for its more complex degradation. The up-to-date dataset for realistic STISR is called TextZoom, while the current methods trained on this dataset have not considered the effect of multi-scale features of text images. In this paper, a multi-scale and attention fusion model for realistic STISR is proposed. The multi-scale learning mechanism is introduced to acquire sophisticated feature representations of text images; The spatial and channel attentions are introduced to capture the local information and inter-channel interaction information of text images; At last, this paper designs a multi-scale residual attention module by skillfully fusing multi-scale learning and attention mechanisms. The experiments on TextZoom demonstrate that the model proposed increases scene text recognition's (ASTER) average recognition accuracy by 1.2% compared to text super-resolution network.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제6권4호
• /
• pp.1188-1202
• /
• 2012

In this paper, we propose a new multi-scale connected coherence tree algorithm (MCCTA) by improving the connected coherence tree algorithm (CCTA). In contrast to many multi-scale image processing algorithms, MCCTA works on multiple scales space of an image and can adaptively change the parameters to capture the coarse and fine level details. Furthermore, we design a Multi-scale Connected Coherence Tree algorithm plus Spectral graph partitioning (MCCTSGP) by combining MCCTA and Spectral graph partitioning in to a new framework. Specifically, the graph nodes are the regions produced by CCTA and the image pixels, and the weights are the affinities between nodes. Then we run a spectral graph partitioning algorithm to partition on the graph which can consider the information both from pixels and regions to improve the quality of segments for providing image segmentation. The experimental results on Berkeley image database demonstrate the accuracy of our algorithm as compared to existing popular methods.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2021년도 추계학술발표대회
• /
• pp.655-658
• /
• 2021

The purpose of multimodal medical image fusion (MMIF) is to integrate images of different modes with different details into a result image with rich information, which is convenient for doctors to accurately diagnose and treat the diseased tissues of patients. Encouraged by this purpose, this paper proposes a novel method based on a two-scale decomposer and detail preservation model. The first step is to use the two-scale decomposer to decompose the source image into the energy layers and structure layers, which have the characteristic of detail preservation. And then, structure tensor operator and max-abs are combined to fuse the structure layers. The detail preservation model is proposed for the fusion of the energy layers, which greatly improves the image performance. The fused image is achieved by summing up the two fused sub-images obtained by the above fusion rules. Experiments demonstrate that the proposed method has superior performance compared with the state-of-the-art fusion methods.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2007년도 Proceedings of ISRS 2007
• /
• pp.452-455
• /
• 2007

3D stereoscopic viewing of large scale imagery, such as aerial photography and satellite images, needs different parallaxes relative to the display scale. For example, when a viewer sees a stereoscopic image of aerial photography, the optimal parallax of its zoom-in image should be smaller than that of its zoom-out. Therefore, relative parallax adjustment according to the display scale is required. Merely adjusting the spacing between stereo images is not appropriate because the depths of the whole image are either exaggerated or reduced entirely. This paper focuses on the improving stereoscopic viewing with a single remote sensing image and a digital surface model (DSM). We present the parallax adjustment technique to maximize the 3D realistic effect and the visual comfort. For remote sensing data, DSM height value can be regarded as disparity. There are two possible kinds of methods to adjust the relative parallax with a single image performance. One is the DSM compression technique: the other is an adjustment of the distance between the original image and its stereo-mate. In our approach, we carried out a test to evaluate the optimal distance between a single remote sensing image and its stereo-mate, relative to the viewing scale. Several synthetic stereo-mates according to certain viewing scale were created using a parallel projection model and their anaglyphs were estimated visually. The occlusion of the synthetic stereo-mate was restored by the inpainting method using the fields of experts (FoE) model. With the experiments using QuickBird imagery, we could obtain stereoscopic images with optimized parallax at varied display scales.

• 패션비즈니스
• /
• 제8권2호
• /
• pp.42-55
• /
• 2004

The purpose of this study was to classify the contents of department stores and large-scale discount stores of consumer on information source, shopping orientation and store image in an effort to determine which variable gave a crucial impact on cosmetics department stores and large-scale discount stores selection behavior. The subjects of this study were 557 adult women visited department store and large-scale discount store in Busan. The data were analyzed by using Factor analysis, Frequency analysis, Correlation analysis, Cronabach $\alpha$ and Regression analysis. The results were as follows; 1. There was a difference in the demographical characteristics on department stores and large-scale discount stores of consumers. 2. Shopping Orientation of consumers were categorized into 5 types, and information source of consumers did 3 types, and store image of consumers did 5 types. 3. Leisure utilization, store & brand loyalty, store information, personal information, massmedia information, product & operate on, store atmosphere & salesperson and shopping convenience had positive correlations with cosmetics department stores selection beavior. but convenient location, rationality & economy and sales promotion had negative correlations with cosmetics department stores selection beavior. 4. Rationality & economy and sales promotion had positive correlations with cosmetics large-scale discount stores selection beavior. but convenient location, leisure utilization, store & brand loyalty, massmedia information, product & operate on, store atmosphere & salesperson and shopping convenience had negative correlations with cosmetics large-scale discount stores selection beavior. 5. Age, income, business(-), convenient location(-), rationality & economy(-), leisure utilization, store & brand loyalty, store information, personal information, massmedia information, store atmosphere & salesperson, shopping convenience and sales promotion(-) had a direct effect on cosmetics department stores selection beavior. Age, income, marriage, education had an indirect effect on department stores selection beavior through information source and store image, and information source did through store image, and shopping orientation did through store image. 6. Rationality & economy, convenient location(-), leisure utilization(-), store & brand loyalty(-), buying independence(-), personal information, massmedia information(-), product & operate on(-), shopping convenience(-) and sales promotion had a direct effect on cosmetics large-scale discount stores selection beavior. Age, income, marriage, education had an indirect effect on large-scale discount stores selection beavior through information source, shopping orientation and store image, and information source did through store image, and shopping orientation did through store image.

• 반도체디스플레이기술학회지
• /
• 제12권3호
• /
• pp.59-64
• /
• 2013

Accurate and robust image registration is important task in many applications such as image retrieval and computer vision. To perform the image registration, essential required steps are needed in the process: feature detection, extraction, matching, and reconstruction of image. In the process of these function, feature extraction not only plays a key role, but also have a big effect on its performance. There are two representative algorithms for extracting image features, which are scale invariant feature transform (SIFT) and speeded up robust feature (SURF). In this paper, we present and evaluate two methods, focusing on comparative analysis of the performance. Experiments for accurate and robust feature detection are shown on various environments such like scale changes, rotation and affine transformation. Experimental trials revealed that SURF algorithm exhibited a significant result in both extracting feature points and matching time, compared to SIFT method.

• 정보과학회 논문지
• /
• 제42권1호
• /
• pp.93-96
• /
• 2015

이미지 검색 및 매칭에 사용되는 SIFT 기법은 다양한 이미지 변화 요인들에 대하여 강인한 특성을 가지고 있는 것으로 알려져 있다. SIFT 기법은 기존의 픽셀 단위의 변화량에 의존한 특징점 추출 방식을 확장하여 스케일 공간에서의 변화량 분석을 통한 특징점 추출 방식을 제시하였으며, 이렇게 추출된 특징점들의 강인함은 그 동안 여러 실험을 통하여 입증되었다. 또한, 최근에는 스케일 공간 변화량 분석에 있어서 기존의 SIFT 기법을 확장하여 고차 미분 계수를 이용한 특징점 추출 방법도 소개되었다. 본 논문에서는 이러한 스케일 공간의 고차 미분에서의 정규화를 통한 보다 강인한 특징점 추출 기법을 소개하고 이러한 특징점들의 강인함을 이미지 검색 실험을 통하여 입증한다.

• 한국가시화정보학회지
• /
• 제1권2호
• /
• pp.13-16
• /
• 2003

LSPIV(Large Scale Particle Image Velocimetry) is widely used in the field of civil and environmental engineering. General aspects of LSPIV are introduced and several applications are introduced in this paper. The difference of LSPIV from the conventional PIV techniques is not to use models for experiments but to use the flow fields in nature. For LSPIV a converting process for the captured images is necessary.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1999년도 추계종합학술대회 논문집
• /
• pp.589-592
• /
• 1999

Matching is a key problem in computer vision, image analysis and pattern recognition. In this paper a multiscale image matching algorithm by wavelet local extrema is proposed. This algorithm is based on the multiscale wavelet transform of the curvature which can utilize both the information of local extrema positions and magnitudes of transform results. This method has advantages in computational cost to a single scale image matching. It is also rotation-, translation-, and scale-independent image matching method. This matching can be used for the recognition of occluded objects.

• 농촌계획
• /
• 제17권2호
• /
• pp.33-41
• /
• 2011

To evaluate the value of a preserved tree as rural landscape resource, the growth environment and health condition was investigated, and the image evaluation was implemented on land~ape architectural major undergraduate students for zelkova trees in Yesan-gun. The image evaluation results of zelkova trees were as followings; 1) Typical image of preserved tree examined by Semantic Differential Scale were 'Old', 'Big', and 'Good'. 2) The 'big' image of zelkova tree and the height of tree, the width of tree crown, the breast girth of tree, the root girth of tree, the external formation of tree, and the health of tree bark is mutually related. Especially, the correlation between the 'big' and the external formation and the width of tree crown is high. 3) Typical image of preserved tree examined by Likert Scale were 'Natural', 'Green', 'Peaceful', and 'Rural'. 4) The preservation necessity for preserved tree was highly related with the state of ground, and the management necessity for preserved tree was highly related with contamination level and the state of ground. The appropriate management plan for preserved tree are proposed to improve the quality of rural landscape(basis of these results).