• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.246-249
• /
• 2003

The constraints necessary guarantee using the comparison of these extrinsic parameters, which each Rotation matrix and Translation Vector must be equal to the either, except the X-axis Translation Vector. Thus, we can not yet calculate the 3D-range measurement in the end of camera calibration. To minimize this disadvantage, the Epipolar Rectification has been proposed in the literature. This paper aims to present the development of Epipolar Rectification to calibrate Stereo cameras. The required computation of the transformation mapping between points in 3D-space is based on calculating the image point that appears on new image plane by using calibrated parameters. This computation is assumed from the rotating the old ones around their optical center until focal planes becomes coplanar, thereby containing the baseline, and the Z-axis of both camera coordinate to be parallel together. The optical center positions of the new extrinsic parameters are the same as the old camera, whereas the new orientation differs from the old ones by the suitable rotations. The intrinsic parameters are the same for both cameras. So that, after completed calibration process, immediately can calculate the 3D-range measurement. And the rectification determines a transformation of each image plane such that pairs of conjugate Epipolar lines become collinear and parallel to one of the image axis. From the experimental results verify the proposed technique are agreed with the expected specifications.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.35 no.6
• /
• pp.603-608
• /
• 2011

Since PTV (particle tracking velocimetry) provides velocity vectors by tracking each particle in a fluid flow, it has significant benefits when used for nano- and bio-fluid flows. However, PTV has only been used for limited flow fields because interpolation data loss is inevitable in PTV in principle. In this paper, a hybrid particle image velocimetry (PIV) algorithm that eliminates interpolation data loss was constructed by using an affine transformation. For the evaluation of the performance of the constructed hybrid PIV algorithm, an artificial image test was performed using Green-Taylor vortex data. The constructed algorithm was tested on experimental images of the wake flow (Re = 5,300) of a rectangular body ($6cm\;{\times}3cm$), and was demonstrated to provide excellent results.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.47-51
• /
• 2005

Using satellite images, an optimal solution to water motion has been presented in this study. Since temperature patterns are suitable tracers in water motion, Sea Surface Temperature (SST) images of Caspian Sea taken by MODIS sensor on board Terra satellite have been used in this study. Two daily SST images with 24 hours time interval are used as input data. Computation of templates correspondence between pairs of images is crucial within motion algorithms using non-rigid body objects. Image matching methods have been applied to estimate water body motion within the two SST images. The least squares matching technique, as a flexible technique for most data matching problems, offers an optimal spatial solution for the motion estimation. The algorithm allows for simultaneous local radiometric correction and local geometrical image orientation estimation. Actually, the correspondence between the two image templates is modeled both geometrically and radiometrically. Geometric component of the model includes six geometric transformation parameters and radiometric component of the model includes two radiometric transformation parameters. Using the algorithm, the parameters are automatically corrected, optimized and assessed iteratively by the least squares algorithm. The method used in this study, has presented more efficient and robust solution compared to the traditional motion estimation schemes.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.1
• /
• pp.11-22
• /
• 1997

The fractal image compression partitions an original image into blocks of equal size and searches a do-main block having self-similarity. This method of compression achieves high compression ratio because it is unnecessary to transmit the additional codebook to receiver and it provides good quality of reconstructed images. In spite of these advantages, this method has a drawback in which encoding time increase due to a complicated linear transformation for determining a similar-domain block. In this paper, a fast fractal image compression method is proposed by decreasing the number of transformation usings AC(alternating current) coefficients of block. The proposed method also has a good quality as compared with the well-known fractal codings. Furthermore, method also has a good quality as apply the video steganography that can conceal an important secret data.

• Proceedings of the Korea Contents Association Conference
• /
• 2003.05a
• /
• pp.376-382
• /
• 2003

Recently, there have been many types of wavelet transformations proposed to remove the noise from an signal and image data By using feature of seperating the noise from the original image the Wavelet transformations can retain the edges of the images The wavelet analysis is complete when the basis function is coded into the wavelet This Thesis describes a method of using wavelet transformation to remove the noise from an image signal. Although the wavelet transformation proposed by Donoho and Johnstone works, it does not reliably remove all the noise from the images. So this thesis propose an algorithm that selected Wavelet Shrinkgae and threshold according to the features of bands and amplitude of noise.

• Proceedings of the Korean Society of Medical Physics Conference
• /
• 2002.09a
• /
• pp.442-445
• /
• 2002

We present an algorithm for automatic anatomically adaptive image enhancement of digital chest radiographs. Chest images were exposed using digital radiography system with a 0.143 mm pixel pitch, l4-bit gray levels, and 3121 ${\times}$ 3121 matrix size. A chest radiograph was automatically divided into two classes (lung field and mediastinum) by using a maximum likelihood method. Each pixel in an image was processed using fuzzy domain transformation and enhancement of both the dynamic range and local gray level variations. The lung fields were enhanced appropriately to visualize effectively vascular tissue, the bronchus, and lung tissue, etc as well as pneumothorax and other lung diseases at the same time with the desired mediastinum enhancement. A prototype implementation of the algorithm is undergoing trials in the clinical routine of radiology department of major Korean hospital.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.7 no.4
• /
• pp.119-131
• /
• 2011

In this paper, we explore the application of 2-D dual-tree discrete wavelet transform (DDWT), which is a directional and redundant transform, for image coding. DDWT main property is a more computationally efficient approach to shift invariance. Also, the DDWT gives much better directional selectivity when filtering multidimensional signals. The dual-tree DWT of a signal is implemented using two critically-sampled DWTs in parallel on the same data. The transform is 2-times expansive because for an N-point signal it gives 2N DWT coefficients. If the filters are designed is a specific way, then the sub-band signals of the upper DWT can be interpreted as the real part of a complex wavelet transform, and sub-band signals of the lower DWT can be interpreted as the imaginary part. The quincunx lattice is a sampling method in image processing. It treats the different directions more homogeneously than the separable two dimensional schemes. Quincunx lattice yields a non separable 2D-wavelet transform, which is also symmetric in both horizontal and vertical direction. And non-separable wavelet transformation can generate sub-images of multiple degrees rotated versions. Therefore, non-separable image processing using DDWT services good performance.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.2
• /
• pp.349-354
• /
• 2023

As the population interested in interior design has been increasing, the global interior market has grown significantly. Global interior companies are developing and providing simulation services for various interior elements. Although wallpaper design is the most important interior element, existing wallpaper design simulation services are difficult to use due to drawbacks such as differences between expected and actual results, long simulation time, and the need for professional skills. We proposed a wallpaper image transformation method for interior design using cycle generative adversarial networks (GAN). The proposed method demonstrates that users can simulate wallpaper design within a short period of time based on interior image data using various types of wallpaper.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.10
• /
• pp.2788-2808
• /
• 2023

Brain tumors are one of the most threatening malignancies for humans. Misdiagnosis of brain tumors can result in false medical intervention, which ultimately reduces a patient's chance of survival. Manual identification and segmentation of brain tumors from Magnetic Resonance Imaging (MRI) scans can be difficult and error-prone because of the great range of tumor tissues that exist in various individuals and the similarity of normal tissues. To overcome this limitation, the Amended Convolutional Neural Network (ACNN) model has been introduced, a unique combination of three techniques that have not been previously explored for brain tumor detection. The three techniques integrated into the ACNN model are image tissue preprocessing using the Kalman Bucy Smoothing Filter to remove noisy pixels from the input, image tissue segmentation using the Isotonic Regressive Image Tissue Segmentation Process, and feature extraction using the Marr Wavelet Transformation. The extracted features are compared with the testing features using a sigmoid activation function in the output layer. The experimental findings show that the suggested model outperforms existing techniques concerning accuracy, precision, sensitivity, dice score, Jaccard index, specificity, Positive Predictive Value, Hausdorff distance, recall, and F1 score. The proposed ACNN model achieved a maximum accuracy of 98.8%, which is higher than other existing models, according to the experimental results.

• Journal of Broadcast Engineering
• /
• v.27 no.1
• /
• pp.56-68
• /
• 2022

In this paper, we propose a compression method for multi-resolution feature maps for VCM. The proposed compression method removes the redundancy between the channels and resolution levels of the multi-resolution feature map through PCA-based transformation. According to each characteristic, the basis vectors and mean vector used for transformation, and the transformation coefficient obtained through the transformation are compressed using a VVC-based coder and DeepCABAC. In order to evaluate performance of the proposed method, the object detection performance was measured for the OpenImageV6 and COCO 2017 validation set, and the BD-rate of MPEG-VCM anchor and feature map compression anchor proposed in this paper was compared using bpp and mAP. As a result of the experiment, the proposed method shows a 25.71% BD-rate performance improvement compared to feature map compression anchor in OpenImageV6. Furthermore, for large objects of the COCO 2017 validation set, the BD-rate performance is improved by up to 43.72% compared to the MPEG-VCM anchor.