• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.207-209
• /
• 2021

In recent years, with the development of artificial intelligence and IoT technology, automation and unmanned technology are in progress in various fields, and the importance of image processing such as object tracking, medical images and object recognition, which are the basis of this, is increasing. In particular, in systems requiring detailed data processing, noise reduction is used as a pre-processing step, but the existing algorithm has a disadvantage that blurring occurs in the filtering process. Therefore, in this paper, we propose a filter algorithm using modified spatial weights to minimize information loss in the filtering process. The proposed algorithm uses mask matching to remove AWGN, and obtains the output of the filter by adding or subtracting the output of the modified spatial weight. The proposed algorithm has superior noise reduction characteristics compared to the existing method and reconstructs the image while minimizing the blurring phenomenon.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.11
• /
• pp.31-37
• /
• 2018

This paper proposes an effective single-image haze-removal algorithm with low complexity by using a dual dark channel prior (DCP) and an adaptive brightness correction technique. The dark channel of a small patch preserves the edge information of the image, but is sensitive to noise and local brightness variations. On the other hand, the dark channel of a large patch is advantageous in estimation of the exact haze value, but halo effects from block effects deteriorate haze-removal performance. In order to solve this problem, the proposed algorithm builds a dual DCP as a combination of dark channels from patches with different sizes, and this meets low-memory and low-complexity requirements, while the conventional method uses a matting technique, which requires a large amount of memory and heavy computations. Moreover, an adaptive brightness correction technique that is applied to the recovered image preserves the objects in the image more clearly. Experimental results for various hazy images demonstrate that the proposed algorithm removes haze effectively, while requiring much fewer computations and less memory than conventional methods.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.7
• /
• pp.36-41
• /
• 2018

An image intensifying device is equipment that makes weak objects visible in a dark environment, such as making nighttime bright enough to let objects be visually observed. It is possible to obtain a clear image by amplifying the light in the presence of a certain amount of weak light. However, in an extremely low-light environment, where even moonlight is not present, there is not enough light to amplify anything, and the sharpness of the screen deteriorates. In this paper, a method is proposed to improve image quality by using multiple filters and anisotropic diffusion for output noise of the image-intensifying device in extreme low-light environments. For the experiment, the output of the image-intensifying device was obtained under extremely low-light conditions, and signal processing for improving the image quality was performed. The configuration of the filters for signal processing uses anisotropic diffusion after applying a median filter and a Wiener filter for effective removal of salt-and-pepper noise and Gaussian noise, which constitute the main noise appearing in the image. Experimental results show that the improvement visually enhanced image quality. Both peak signal-to-noise ratio (PSNR) and SSIM, which are quantitative indicators, show improved values.

• KIISE Transactions on Computing Practices
• /
• v.20 no.11
• /
• pp.573-578
• /
• 2014

In this letter, we propose a noise reduction method for use in magnetic resonance images that is based on non-local mean and guided image filters. Our method consists of two phases. In the first phase, the guidance image is obtained from a noisy image by using an adaptive non-local mean filter. The spread of the kernel is adaptively by controlled by implementing the concept of edgeness. In the second phase, the noisy images and the guidance images are provided to the guided image filter as input in order to produce a noise-free image. The improved performance of the proposed method is investigated by conducting experiments on standard datasets that contain magnetic resonance images. The results show that the proposed scheme is superior over the existing approaches.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.8A
• /
• pp.679-688
• /
• 2012

In this paper, performance improved hole-filling algorithm including boundary noise removing pre-process which can be used for an arbitrary view synthesis with given two views is proposed. Boundary noise usually occurs because of the boundary mismatch between the reference image and depth map and common-hole is defined as the occluded region. These boundary noise and common-hole created while synthesizing a virtual view result in some defects and they are usually very difficult to be completely recovered by using only given two images as references. The spiral weighted average algorithm gives a clear boundary of each object by using depth information and the gradient searching algorithm is able to preserve details. In this paper, we combine these two algorithms by using a weighting factor ${\alpha}$ to reflect the strong point of each algorithm effectively in the virtual view synthesis process. The experimental results show that the proposed algorithm performs much better than conventional algorithms.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.22 no.8
• /
• pp.1090-1098
• /
• 2018

Many short range weather radars with the low elevation search capability are needed for analysis and prediction of unusual weather changes or rainfall phenomena which occurs regionally. However, due to the characteristics of low elevation electromagnetic wave beam, it is highly probable that the received weather signals of these radars are seriously contaminated by the ground clutter. Therefore, the filter removing low Doppler frequency band is generally used to mitigate this problem. However, the phase noise in a radar system may limit the removal of the strong clutter and this may cause serious problems in estimating weather parameters because of the remaining clutter. Therefore, in this paper, the characteristics of phase noise in a radar system are investigated and the effects of the system phase noise are analyzed in the improvement of signal to clutter ratio for the strong clutter environment such as a short and low-elevated weather radar.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.11
• /
• pp.1637-1643
• /
• 2022

In recent years, the demand for image-processing technology in digital marketing has increased due to the expansion and diversification of the digital market, such as video, security, and machine intelligence. Noise-processing is essential for image-correction and reconstruction, especially in the case of sensitive noises, such as in CT, MRI, X-ray, and scanners. The two main salt and pepper noises have been actively studied, but the details and edges are still unsatisfactory and tend to blur when there is a lot of noise. Therefore, this paper proposes an algorithm that applies a weight-based Euclidean distance equation to the partial mask and uses only the non-noisy pixels that are the most similar to the original as effective pixels. The proposed algorithm determines the type of filter based on the state of the internal pixels of the designed partial mask and the degree of mask deterioration, which results in superior noise cancellation even in highly damaged environments.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.6
• /
• pp.107-116
• /
• 2010

Smear is a phenomenon that occurs when an extremely strong light source appears in the imaging system with CCD sensor. It occurs due to the signal charge transfer of CCD and appears as bright lines of noise emanating vertically (or horizontally) from the light source. For still images, smear can be reduced by using a mechanical shutter or special drive methods, but these techniques cannot be applied to image sequences. In this paper, we propose a smear removal method that can be applied to imaging systems for not only still images but also image sequences. The proposed method uses the optical black region(OBR) which is a group of pixels located in the boundary of CCD imaging sensors. Although the OBR is not exposed to light, it contains smear information caused by the charge transport. First, noise and the smear signal in the OBR is separated, and noise is removed to correctly estimate smear effect. Then, corrected OBR signal is uniformly subtracted to eliminate smear effect. Also, if saturation is occurred, the current pixel is substituted by weighted summation of neighboring pixels to improve the visual degradation. Experimental results show that the proposed algorithm outperforms the conventional methods.

• Journal of IKEEE
• /
• v.25 no.2
• /
• pp.273-279
• /
• 2021

In this paper, we propose an algorithm that shortens the learning time by performing individual learning using partitioning the deep learning structure. The proposed algorithm consists of four processes: network classification origin setting process, feature vector extraction process, feature noise removal process, and class classification process. First, in the process of setting the network classification starting point, the division starting point of the network structure for effective feature vector extraction is set. Second, in the feature vector extraction process, feature vectors are extracted without additional learning using the weights previously learned. Third, in the feature noise removal process, the extracted feature vector is received and the output value of each class is learned to remove noise from the data. Fourth, in the class classification process, the noise-removed feature vector is input to the multi-layer perceptron structure, and the result is output and learned. To evaluate the performance of the proposed algorithm, we experimented with the Extended Yale B face database. As a result of the experiment, in the case of the time required for one-time learning, the proposed algorithm reduced 40.7% based on the existing algorithm. In addition, the number of learning up to the target recognition rate was shortened compared with the existing algorithm. Through the experimental results, it was confirmed that the one-time learning time and the total learning time were reduced and improved over the existing algorithm.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.3
• /
• pp.129-138
• /
• 2013

This paper is a study on Dual Tree Complex Wavelet Transform, which improved directional information for efficient image processing. Dual Tree Complex Wavelet Transform satisfies characteristics of shift invariance, and includes 6 directional information, which is more than previous Discrete Wavelet Transform. However, in images of buildings, there are many horizontal and vertical edge components. Therefore, all the high-frequency components of image are not expressed by 6 directional information subbands. This paper proposes 8-directional Complex Wavelet Transform with excellent high-frequency separation features by creating horizontal vertical($0^{\circ}$, $90^{\circ}$) subband besides 6 directional information subband of previous Dual Tree Complex Wavelet Transform. The proposed method can create and combine various directional information subbands according to features of image. Performance is evaluated by applying the method to noise removal.