• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.2
• /
• pp.200-208
• /
• 2003

This paper proposes a separation and feature extraction of image signals using the independent component analysis(ICA) based on neural networks of efficient learning rule. The proposed learning rule is a hybrid fixed-point(FP) algorithm based on secant method and momentum. Secant method is applied to improve the performance by simplifying the 1st-order derivative computation for optimizing the objective function, which is to minimize the mutual informations of the independent components. The momentum is applied for high-speed convergence by restraining the oscillation in the process of converging to the optimal solution. The proposed algorithm has been applied to the composite images generated by random mixing matrix from the 10 images of $512\times512$-pixel. The simulation results show that the proposed algorithm has better performances of the separation speed and rate than those using the FP algorithm based on Newton and secant method. The proposed algorithm has been also applied to extract the features using a 3 set of 10,000 image patches from the 10 fingerprints of $256\times256$-pixel and the front and the rear paper money of $480\times225$-pixel, respectively, The simulation results show that the proposed algorithm has also better extraction speed than those using the another methods. Especially, the 160 basis vectors(features) of $16\times16$-pixel show the local features which have the characteristics of spatial frequency and oriented edges in the images.

• The Journal of the Korea Contents Association
• /
• v.9 no.4
• /
• pp.189-197
• /
• 2009

A digital image usually has 8 bits of depth basically representing pixel intensity ranging for [0 255]. These pixel range allow 256 step levels of pixel values in the image. Thus the greyscale value for a given image is an integer. When we carry out interpolation of a given image for resizing we have to round the interpolated value to integer which can result in loss of quality on perceived color values. This paper proposes a new method for recovering this loss of information during interpolation process. By using the proposed method the pixels tend to regain more original values which yields better looking images on resizing.

• Journal of Sensor Science and Technology
• /
• v.10 no.6
• /
• pp.317-327
• /
• 2001

In this paper, we proposed a new detection method of pH image to effectively measure a 2-dimensional pH distribution of test materials by irradiating an frequency modulated light to LAPS using a reticle. It could measure simultaneously signals in one line by applying a modulated light having difference frequency for each pixel using a frequency modulating reticle, and calculating an amplitude with respect to a frequency component by the light source. To experiment the proposed method, we designed and implemented a reticle considering of a LAPS's characteristic, and reconstructed an image by frequency analysis using the implemented reticle and test pattern image. As a result, we verified that the proposed method using the reticle was able to detect 30 times faster for a $30{\times}30$ pixels pH image having a PSNR of 22-24 [dB] than conventional method.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.12 no.5
• /
• pp.205-211
• /
• 2012

By using an adaptive threshold and weight based on the wavelet transform and Haar transform, a novel image enhancement algorithm is proposed. First, a medical image was decomposed with wavelet transform and all high-frequency sub-images were decomposed with Haar transform. Secondly, noise in the frequency domain was reduced by the proposed soft-threshold method. Thirdly, high-frequency coefficients were enhanced by the proposed weight values in different sub-images. Then, the enhanced image was obtained through the inverse Haar transform and wavelet transform. But the pixel range of the enhanced image is narrower than a normal image. Lastly, the image's histogram was stretched by nonlinear histogram equalization. Experiments showed that the proposed method can be not only enhance an image's details but can also preserve its edge features effectively.

• Journal of Broadcast Engineering
• /
• v.16 no.3
• /
• pp.552-560
• /
• 2011

In this paper, a new no-reference sharpness measure using frequency domain coefficients is proposed. Although most existing sharpness measures used pixel intensity to compute the blur degree, the proposed sharpness measure computes the sharpness using frequency coefficients. To assess the perceived sharpness of a given image, the image is re-blurred by a Gaussian low pass filter and a new quality measure function was defined using the frequency domain coefficients of the given image and the re-blurred image. To evaluate the proposed algorithms, TID2008 quality assessment database was used. Experimental results show that the proposed quality assessment method showed high correlation with the subjective scores.

• The Korean Journal of Applied Statistics
• /
• v.16 no.2
• /
• pp.305-319
• /
• 2003

This paper provides a non-Bayesian method based on the expanded EM algorithm for segmenting the magnetic resonance images degraded by bias field. For the images with the intensity as a pixel value, many segmentation methods often fail to segment it because of the bias field(with low frequency) as well as noise(with high frequency). Our contextual approach is appropriately designed by using normal mixture model incorporated with Markov random field for noise-corrective segmentation and by using the penalized likelihood to estimate bias field for efficient bias filed-correction.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.755-758
• /
• 2001

In this paper. Ive propose a postprocessing technique based on the theory of projection on convex sets(POCS) to reduce the blocking artifacts in HDTV decoded images. In BDCT of HDTV. the image is divided into a grid of non-overlapped 8 ${\times}$ 8 blocks. and then each block is coded separately. A block which is located one pixel apart from the grid of BDCT will include the boundary of the original 8 ${\times}$ 8 block. If the blocking artifact is Introduced alone the block boundary. this block will have different frequency characteristic from that of the original block. Thus, a comparison of frequency characteristics of these two overlapping blocks can detect the undesired high-frequency components mainly caused by the blocking artifact. By eliminating these undesired high-frequency components adaptively, robust smoothing projection operator can be obtained. Simulation results with real image sequences indicate that the proposed method performs better than conventional algorithms.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.255-257
• /
• 2018

Recently, as the frequency of use of video media increases in various fields, the importance of signal processing is increasing. However, many kinds of noise are generated in the transmission and reception process and affect the information of the signal. For this reason, the noise removal is essential as a preprocessing process. In this paper, we propose an algorithm to remove mixed noise of impulse noise and AWGN. The proposed algorithm restores the image through noise determination and pixel change for efficient noise removal. Unlike the conventional method, noise is removed by minimizing both noise effects. Simulation showed excellent noise removal characteristic results were compared and analyzed using the PSNR for such decisions.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.44 no.3
• /
• pp.29-33
• /
• 2007

All pixels of image sensor do not react uniformly even if the light of same radiance enters into the camera. This non-uniformity comes from the sensor pixel non-uniformity and non-uniformity induced by the changing transmission of the telescope over the field. The first contribution to the non-uniformity has high spatial frequency nature and has an influence on the result and quality of the data compression. The second source of non-uniformity has low frequency nature and has no influence of the compression result. As the contribution resulting from the sensor PRNU(Photo Response Non-Uniformity) is corrected inside the camera electronics, the effect of the remaining non-uniformity to the compression result will be negligible. The non-uniformity correction result shall have big difference according to the sensor modeling and the calculation method to get correction coefficient. Usually, the sensor can be modeled with one dimensional coefficients which are a gain and a offset for each pixel. Only two measurements are necessary theoretically to get coefficients. However, these are not the optimized value over the whole illumination level. This paper proposes the algorithm to calculate the optimized non-uniformity correction coefficients over whole illumination radiance. The proposed algorithm uses several measurements and the least square method to get the optimum coefficients. The proposed algorithm is verified using the own camera electronics including sensor, electrical test equipment and optical test equipment such as the integrating sphere.

• Proceedings of the IEEK Conference
• /
• 2000.06d
• /
• pp.210-213
• /
• 2000

In the closed range space, the parallel two CCD cameras are used to acquire a pair of stereo image. The acquired stereo image are computed with Wavelet Transform repeatedly and including the low frequency component, the image size of those are reduced. It is the pyramid structure. The optimum matching point is searched to the pixel. Then appling the optimum matching point to DLT, it extract the three - dimensional surface coordinate from a stereo image. The direct linear transformation(DLT) method is used to calibrate the stereo camera compute the coordinate on a three dimensional space. To find the parameters for the DLT method, 30 control points which marked on the cylinder type object are used. To improve the matching algorithm, the paper select the pyramid structure for Wavelet Transform. The acquired disparity information is used to represent the really three-dimensional surface coordinate.