• Proceedings of the IEEK Conference
• /
• 2002.07a
• /
• pp.172-175
• /
• 2002

This paper presents perceptual model with a stochastic rnultiresolution characteristic that can be applied with watermark embedding in the biorthogonal wavelet domain. The perceptual model with adaptive watermarking algorithm embed at the texture and edge region for more strongly embedded watermark by the SSQ(successive subband quantization). The watermark embedding is based on the computation of a NVF(noise visibility function) that have local image properties. This method uses non-stationary Gaussian model stationary Generalized Gaussian model because watermark has noise properties. In order to determine the optimal NVF, we consider the watermark as noise. The particularities of embedding in the stationary GG model use shape parameter and variance of each subband regions in multiresolution. To estimate the shape parameter, we use a moment matching method. Non-stationary Gaussian model use the local mean and variance of each subband. The experiment results of simulation were found to be excellent invisibility and robustness. Experiments of such distortion are executed by Stirmark benchmark test.

• Journal of the Optical Society of Korea
• /
• v.17 no.2
• /
• pp.213-218
• /
• 2013

Although infrared focal plane array (IRFPA) detectors have been commonly used, non-uniformity correction (NUC) remains an important problem in the infrared imaging realm. Non-uniformity severely degrades image quality and affects radiometric accuracy in infrared imaging applications. Residual non-uniformity (RNU) significantly affects the detection range of infrared surveillance and reconnaissance systems. More effort should be exerted to improve IRFPA uniformity. A novel NUC method that considers the surrounding temperature variation compensation is proposed based on the binary nonlinear non-uniformity theory model. The implementing procedure is described in detail. This approach simultaneously corrects response nonlinearity and compensates for the influence of surrounding temperature shift. Both qualitative evaluation and quantitative test comparison are performed among several correction technologies. The experimental result shows that the residual non-uniformity, which is corrected by the proposed method, is steady at approximately 0.02 percentage points within the target temperature range of 283 K to 373 K. Real-time imaging shows that the proposed method improves image quality better than traditional techniques.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.586-588
• /
• 2005

The MSC (Multi Spectral Camera) system is a remote sensing payload to obtain high resolution ground image. This system uses lossy image compression method for &Direct mission& that transmit whole image during one contact. But some image degradation occurred especially at high compression ratio. To reduce this degradation, the MSC uses NUC (Non-uniformity Correction) Unit. This unit correct CCD (Charge Coupled Device)'s high-frequency non-uniformity. So high frequency contents of image can be minimized and whole system SNR can be maximized. But NUC has some disadvantage either. It decreases entire system reliability by adding one electronic system. Adding NUC also led to difficulty of electronic design, assembly and testability. In this paper, the comparison is performed between on-orbit non-uniform correction and on ground correction. by evaluating NUC advantage for the point of view of image quality. Using real MSC parameter and proper model, considerable reference point for the system design came to possible.

• Proceedings of the IEEK Conference
• /
• 2007.07a
• /
• pp.249-250
• /
• 2007

Background subtraction is a method typically used to segment moving regions in image sequences taken from a static camera by comparing each new frame to a model of the scene background. We present a improved non-parametric background model by null hypothesis. Evaluation shows that this approach achieves very sensitive detection with very low false alarm rates.

• Journal of the Korea Society for Simulation
• /
• v.29 no.4
• /
• pp.21-31
• /
• 2020

NCTR(Non-Cooperative Target Recognition) refers to the function of radar to identify target on its own without support from other systems such as ELINT(ELectronic INTelligence). ISAR(Inverse Synthetic Aperture Radar) image is one of the representative methods of NCTR, but it is difficult to automatically classify the target without an identification database due to the significant changes in the image depending on the target's maneuver and location. In this study, we discuss how to build an identification database using simulation and deep-learning technique even when actual images are insufficient. To simulate ISAR images changing with various radar operating environment, A model that generates and learns images through the process named 'Perfect scattering image,' 'Lost scattering image' and 'JEM noise added image' is proposed. And the learning outcomes of this model show that not only simulation images of similar shapes but also actual ISAR images that were first entered can be classified.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.27 no.8
• /
• pp.758-765
• /
• 2016

This paper proposes a novel method for the recognition of the inverse synthetic aperture radar(ISAR) image of multiple targets flying in formation. Rather than separating the ISAR image of each target, the proposed method combines an ISAR image obtained by fusing the ISAR images in the training database. Fusion is conducted by optimizing the non-linear problem whose parameters are the aspect angle and the target location. Assuming that the aspect angle is properly estimated, the proposed method estimates the number of the targets and their locations by optimizing the template matching using PSO. In simulations using the F-16 scale model, the efficiency of the proposed method was demonstrated by yielding the ISAR image identical to that of targets in formation.

• Journal of the Korean Graphic Arts Communication Society
• /
• v.25 no.1
• /
• pp.65-80
• /
• 2007

The input and output device requires precise color representation and CMS (Color Management System) because of the increasing number of ways to apply the digital image into electronic publishing. However, there are slight differences in the device dependent color signal among the input and output devices. Also, because of the non-linear conversion of the input signal value to the output signal value, there are color differences between the original copy and the output copy. It seems necessary for device-dependent color information values to change into device-independent color information values. When creating an original copy through electronic publishing, there should be color management with the input and output devices. From the devices' three phases of calibration, characterization and color conversion, the device-dependent color should undergo a color transformation into a device-independent color. In this paper, an experiment was done where the input device used the linear multiple regression and the sRGB color space to perform a color transformation. The output device used the GOG, GOGO and sRGB for the color transformation. After undergoing a color transformation in the input and output devices, the best results were created when the original target underwent a color transformation by the scanner and digital camera input device by the linear multiple regression, and the LCD output device underwent a color transformation by the GOG model.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.8 no.8
• /
• pp.1129-1135
• /
• 2013

A new vehicle detection algorithm based on the multiple optimal Adaboost classifiers with optimal feature selection is proposed. It consists of two major modules: 1) Theoretical DDISF(Distance Dependent Image Scaling Factor) based image scaling by site modeling of the installed cameras. and 2) optimal features selection by Haar-like feature analysis depending on the distance of the vehicles. The experimental results of the proposed algorithm shows improved recognition rate compare to the previous methods for vehicles and non-vehicles. The proposed algorithm shows about 96.43% detection rate and about 3.77% false alarm rate. These are 3.69% and 1.28% improvement compared to the standard Adaboost algorithmt.

• Journal of Biomedical Engineering Research
• /
• v.36 no.5
• /
• pp.211-220
• /
• 2015

Recently, model-based iterative reconstruction (MBIR) has played an important role in transmission tomography by significantly improving the quality of reconstructed images for low-dose scans. MBIR is based on the penalized-likelihood (PL) approach, where the penalty term (also known as the regularizer) stabilizes the unstable likelihood term, thereby suppressing the noise. In this work we further improve MBIR by using a more expressive regularizer which can restore the underlying image more accurately. Here we used a spline regularizer derived from a linear combination of the two-dimensional splines with first- and second-order spatial derivatives and applied it to a non-quadratic convex penalty function. To derive a PL algorithm with the spline regularizer, we used a separable paraboloidal surrogates algorithm for convex optimization. The experimental results demonstrate that our regularization method improves reconstruction accuracy in terms of both regional percentage error and contrast recovery coefficient by restoring smooth edges as well as sharp edges more accurately.

• Proceedings of the KIEE Conference
• /
• 1999.07g
• /
• pp.3048-3050
• /
• 1999

This paper suggests a position estimating algorithm using mono vision system with projective geometry method. Generally, 3-D information can not be easily extracted from mono vision system which is taken by a camera at a specific point. But this defect is overcome by adopting model-based image analysis and selecting lines and points on the ground as natural landmarks. And this paper suggests a method that estimates position from many natural landmarks by distance error weight function.