• Journal of IKEEE
• /
• v.28 no.3
• /
• pp.390-396
• /
• 2024

Vision Transformer (ViT) learns relationships between patches, but it may overlook important features such as color, texture, and boundaries, which can result in performance limitations in fields like medical imaging or facial recognition. To address this issue, this study proposes the Pairwise Attention Reinforcement (PAR) model. The PAR model takes both the training image and a reference image as input into the encoder, calculates the similarity between the two images, and matches the attention score maps of images with high similarity, reinforcing the matching areas of the training image. This process emphasizes important features between images and allows even subtle differences to be distinguished. In experiments using clock-drawing test data, the PAR model achieved a Precision of 0.9516, Recall of 0.8883, F1-Score of 0.9166, and an Accuracy of 92.93%. The proposed model showed a 12% performance improvement compared to API-Net, which uses the pairwise attention approach, and demonstrated a 2% performance improvement over the ViT model.

• Cartoon and Animation Studies
• /
• s.9
• /
• pp.127-142
• /
• 2005

Not only in Korea but throughout the entire world millions of people are in contact with images. Images have become a medium through which to transmit anything from simple visualizations of moving images to knowledge and information. The age of the internet has arisen thanks to scientific development, and the internet generation's acquisition of information is continuously becoming faster. The spectators, ufo must choose amongst the excessive amount of available information, are changing along with it just as quickly. The method of visual transmission has changed to match the demands of the fast-changing pace of the new generation. In order to receive an instantaneous selection amongst much information, the primary requisite is attracting one's attention, and then presenting a corresponding feeling of satisfaction. The early stages of film arose from the desire to capture one's actual situation as it realty is. Unsatisfied with the still picture, people developed the motion picture. Research has succeeded in reproducing 3-dimensional images more realistic than the actual image we perceive as a result of the difference in visual perspective of both eyes and their response to rays of light From color film to 3-dimensional pictures, people enjoy the magnificent results of this. All fields within the category of film are continuously studying the human desire to pursue their visual side, namely the pursuit of visual images with a maximum sense of reality. The images that millions of people around the world see now are flat. The screen's depth and optical illusions effectively give a sense of reality while conveying information. However, although the flat screen is able to create a sense of depth using the different visual perspective of each eye for the realization of a cubic effect, there are limitations. Entering the 21s1 century, there is a quickly-arising branch within the field of image media which seeks to overcome these limitations Although 3-dimensional images began in films, entering the latter half of the 20th century, due to development of 3-dimensional images using the mediums of the animation field, cellular phones, advertisement screens, television etc., without restriction is designated as 'image.'. With research having started around 1900 and continuing for over 100 years, we are now able to witness the popularization of 3-dimensional films happening before our very eyes. Within our own country, we can frequently see them at amusement parks and museums. In the future, through the popularization of HDTV etc., there is a good outlook for practical use of 3-dimensional images in televisions with advanced picture qualify as well as in other areas. Together with the international current, research on 3-dimensional films has been activated in Korea and is rising as a main current in the film industry. Within this context, the contents and understanding of 3-dimensional images must keep in step with the pace of technical advancements. In order to accelerate of development of film contents to keep in pace with technical developments, this dissertation presents the techniques and technical aspects of future developments, and shows the need to prepare in advance to make the field grow- and thereby avoid having a lack of experts and being conquered by other nations in the field - rather than only advancing the technical aspects and importing the contents. This dissertation aims to stimulate interest and continual research by progressive-thinking people related to the film industry. Part II looks into the definition and types of 3-dimensional motion pictures, the terminology, the fundamentals of image formation, current market fluctuations, and looks into 3-dimensional techniques which can be borrowed and introduced in 3-dimensional animations. Part III concerns 3-dimensional animated films. It analyzes 3-dimensional production techniques while using the introduction of specific animation techniques in the 2004 production Lee Sun Shin and Nelson - Naval Heroes 3-dimensional animation produced in 2004 by Clay & Puppet Stop-Motion Animation & Computer Graphic. Original Korean title: 해전영웅 이순신과 넬슨. as an example, and it also looks into how current film techniques used in animations can be applied in 3-dimensional films. Additionally, the actual stages of the various fields of 3-dimensional animations are presented. Given the current direction and advancement of 3-dimensional films making use of animations and the possible realization of this field, the author plans to weigh the development of this yet unexploited new market Not looking at the current progress of the field, but rather the direction of the hypothetical types of animation techniques, the author predicts the marketability and possibility of development of each area.

• Investigative Magnetic Resonance Imaging
• /
• v.1 no.1
• /
• pp.119-124
• /
• 1997

Purpose: To assess the utility of magnetic resonance(MR) cerebral blood volume (CBV) map in the evaluation of brain tumors. Materials and Methods: We performed perfusion MR imaing preoperatively in the consecutive IS patients with intracranial masses(3 meningiomas, 2 glioblastoma multiformes, 3 low grade gliomas, 1 lymphoma, 1 germinoma, 1 neurocytoma, 1 metastasis, 2 abscesses, 1 radionecrosis). The average age of the patients was 42 years (22yr -68yr), composed of 10 males and S females. All MR images were obtained at l.ST imager(Signa, CE Medical Systems, Milwaukee, Wisconsin). The regional CBV map was obtained on the theoretical basis of susceptibility difference induced by first pass circulation of contrast media. (contrast media: IScc of gadopentate dimeglumine, about 2ml/sec by hand, starting at 10 second after first baseline scan). For each patient, a total of 480 images (6 slices, 80 images/slice in 160 sec) were obtained by using gradient echo(CE) single shot echo-planar image(EPI) sequence (TR 2000ms, TE SOms, flip angle $90^{\circ}$, FOV $240{\times}240mm,{\;}matrix{\;}128{\times}128$, slice-thick/gap S/2.S). After data collection, the raw data were transferred to CE workstation and rCBV maps were generated from the numerical integration of ${\Delta}R2^{*} on a voxel by voxel basis, with home made software (${\Delta}R2^{*}=-ln (S/SO)/TE). For easy visual interpretation, relative RCB color coding with reference to the normal white matter was applied and color rCBV maps were obtained. The findings of perfusion MR image were retrospectively correlated with Cd-enhanced images with focus on the degree and extent of perfusion and contrast enhancement. Results: Two cases of glioblastoma multiforme with rim enhancement on Cd-enhanced Tl weighted image showed increased perfusion in the peripheral rim and decreased perfusion in the central necrosis portion. The low grade gliomas appeared as a low perfusion area with poorly defined margin. In 2 cases of brain abscess, the degree of perfusion was similar to that of the normal white matter in the peripheral enhancing rim and was low in the central portion. All meningiomas showed diffuse homogeneous increased perfusion of moderate or high degree. One each of lymphoma and germinoma showed homogenously decreased perfusion with well defined margin. The central neurocytoma showed multifocal increased perfusion areas of moderate or high degree. A few nodules of the multiple metastasis showed increased perfusion of moderate degree. One radionecrosis revealed multiple foci of increased perfusion within the area of decreased perfusion. Conclusion: The rCBV map appears to correlate well with the perfusion state of brain tumor, and may be helpful in discrimination between low grade and high grade gliomas. The further study is needed to clarify the role of perfusion MR image in the evaluation of brain tumor.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.35 no.3
• /
• pp.265-272
• /
• 2002

Spatial distribution of Phytoplankton Pigment Concentration (PPC) and Sea Surface Temperature (SST) around the East Korean Warm Current (EKWC) was described, using both Coastal Zone Color Scanner (CZCS) images and Advanced Very High Resolution Radiometer (AVHRR) images in May, 1980. Water mass in this region can be classified into five categories in the horizontal profile of PPC and SST, nLw (normalized water-leaving radiance) images: (1) coastal cold water region associated with concentrations of dissolved organic material or yellow colored substances and suspended sediments, (2) cold water region of thermal frontal occurred by a combination of phytoplankton absorption and suspended materials, (3) warm water overlay region by the phytoplankton absorption than the suspended materials; (4) warm water region occurred by the low phytoplankton absorption, and (5) offshore region occurred by the high phytoplankton absorption. In particular, the highest PPC (>2.0 mg/m^3) area appeared in the CZCS and AVHRR images with a band shaped distribution of the thermal front and ocean color front region, which is located the coastal cold waters alonB western thermal front of the warm streamer of the EKWC. In this region, the highest PPC occurred by a combination of the high absorption of the phytoplankton (443 nm) and highest reflectance of suspended materials (550 nm). Another high PPC ($\simeq$$6\;mg/m^3$) appeared in the warm water overlay region inside warm streamer. High phytoplankton pigment concentration of this region was corresponding to the short wavelength of 443 nm, which represented phytoplankton absorption of the CZCS image.

• Journal of the Korean Society for information Management
• /
• v.21 no.4 s.54
• /
• pp.187-208
• /
• 2004

In this study, aesthetic impressions were used for a high-level feature of image retrieval. The term, 'aesthetic' has been studied in psychology, art, and literature. It means unconscious, instantaneous parts of visual perception and emotion. The literatures related to aesthetic impressions were reviewed and four kinds of aesthetic impressions were defined operationally : strong impression, soft impression, courteous impression, and refined impression. 66 image files of paintings were sampled randomly from 1100 paintings and low-level color features were extracted from them by a using perceptual color model(Lai, & Tait, 1998). The high-level features of an image, that is, four kinds of aesthetic impressions of each painting were measured by 4 subjects and averaged. In CBIR, 2 subjects performed image retrievals using example queries. They were asked to retrieve images by using the aesthetic impressions or the keywords. In evaluations, subjects showed that they were satisfied with the aesthetic impression-based image retrieval system on the average. And R-precision of the image retrieval with both color features and aesthetic impressions was higher than that of the image retrieval with color features only. But further studies with larger test collections and query sets should be followed for generalization of the result of this study.

• Journal of KIISE:Software and Applications
• /
• v.29 no.11
• /
• pp.809-817
• /
• 2002

This paper describes a new face detection method that is a pre-processing algorithm for on-line face recognition. To complement the weakness of using only edge or rotor features from previous face detection method, we propose the two types of face detection method. The one is a combined method with edge and color features and the other is a center area color sampling method. To prevent connecting the people's face area and the background area, which have same colors, we propose a new adaptive edge detection algorithm firstly. The adaptive edge detection algorithm is robust to illumination variance so that it extracts lots of edges and breakouts edges steadily in border between background and face areas. Because of strong edge detection, face area appears one or multi regions. We can merge these isolated regions using color information and get the final face area as a MBR (Minimum Bounding Rectangle) form. If the size of final face area is under or upper threshold, color sampling method in center area from input image is used to detect new face area. To evaluate the proposed method, we have experimented with 2,100 face images. A high face detection rate of 96.3% has been obtained.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.4
• /
• pp.40-48
• /
• 2012

This paper models the hue shift phenomenon and proposes a hue correction method to give perceptual matching between projector with and without additional white channel. To quantify the hue shift phenomenon for whole hue angle, 24 color patches with the same lightness are frist created along equally-spaced hue angle, and these are displayed one by one both displays with different luminance levels. Next, each hue value of the patches appeared on the projector with additional white channel is adjusted by observers until the hue values of patches on both displays appear the same visually. After obtaining the hue shift values from the color matching experiment, these values are piecewise fit into six polynomial functions, which approximately determine shifted hue amounts for an arbitrary hue values of each pixel in projector with additional white channel and are utilized to correct them. Actually, an input RGB image is converted to CIELAB LCH color space to get hue values of each pixel and this hue value is shifted as much as the amount calculated by the functions of hue shift model for correction. Finally, corrected image is inversely converted to an output RGB image. For an evaluation, the matching experiment with several test images and the z-score comparisons were performed.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.1
• /
• pp.44-49
• /
• 2016

This paper proposed the method to improve performance of a pixel, which has low accuracy, by applying image segmentation methods based on color and multi-threshold of brightness. Stereo matching is the process to find the corresponding point on the right image with the point on the left image. For this process, distance(depth) information in stereo images is calculated. However, in the case of a region which has textureless, stereo matching has low accuracy and bad pixels occur on the disparity map. In the proposed method, the relationship between adjacent pixels is considered for compensating bad pixels. Generally, the object has similar color and brightness. Therefore, by considering the relationship between regions based on segmented regions by means of color and multi-threshold of brightness respectively, the region which is considered as parts of same object is re-segmented. According to relationship information of segmented sets of pixels, bad pixels in the disparity map are compensated efficiently. By applying the proposed method, the results show a decrease of nearly 28% in the number of bad pixels of the image applied the method which is established.

• Journal of KIISE:Software and Applications
• /
• v.27 no.1
• /
• pp.62-73
• /
• 2000

In this paper, we propose a criteria which can be applied to classify conventional color feature based Content Based Image Retrieval (CBIR) methods with its application areas, and a new image retrieval method which can represent sufficient spatial information in the image and is powerful in invariant searching to translation, rotation and enlargement transform. As the conventional color feature based CBIR methods can not sufficiently include the spatial information in the image, in general, they have drawbacks, which are weak to the translation or rotation, enlargement transform. To solve it, they have represented the spatial information by partitioning the image. Retrieval efficiency, however, is decreased rapidly as increasing the number of the feature vectors. We classify conventional methods to ones using 1st order relations and ones using 2nd order relations as their color object relation, and propose a new method using 3rd order relation of color objects which is good for the translation, rotation and enlargement transform. It makes quantized 24 buckets and selects 3 high scored histogram buckets and calculates 3 mean positions of pixels in 3 buckets and 3 angles. Then, it uses them as feature vectors of a given image. Experiments show that the proposed method is especially good at enlarged images and effective for its small calculation.

• Journal of Internet Computing and Services
• /
• v.7 no.2
• /
• pp.23-35
• /
• 2006

This work presents o novel method which extract facial region und features from motion picture automatically and controls the 3D facial expression in real time. To txtract facial region and facial feature points from each color frame of motion pictures a new nonparametric skin color model is proposed rather than using parametric skin color model. Conventionally used parametric skin color models, which presents facial distribution as gaussian-type, have lack of robustness for varying lighting conditions. Thus it needs additional work to extract exact facial region from face images. To resolve the limitation of current skin color model, we exploit the Hue-Tint chrominance components and represent the skin chrominance distribution as a linear function, which can reduce error for detecting facial region. Moreover, the minimal facial feature positions detected by the proposed skin model are adjusted by using edge information of the detected facial region along with the proportions of the face. To produce the realistic facial expression, we adopt Water's linear muscle model and apply the extended version of Water's muscles to variation of the facial features of the 3D face. The experiments show that the proposed approach efficiently detects facial feature points and naturally controls the facial expression of the 3D face model.