• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.7
• /
• pp.3438-3454
• /
• 2018

In this paper, we propose a robust blind watermarking scheme for high-definition video. In the embedding process, luminance component of each frame is transformed by 2-dimensional fast Fourier transform (2D FFT). A secret key is used to generate a matrix of random numbers for the security of watermark information. The matrix is transformed by inverse steerable pyramid transform (SPT). We embed the watermark into the low and mid-frequency of 2D FFT coefficients with the transformed matrix. In the extraction process, the 2D FFT coefficients of each frame and the transformed matrix are transformed by SPT respectively, to produce two oriented sub-bands. We extract the watermark from each frame by cross-correlating two oriented sub-bands. If a video is degraded by some attacks, the watermarks of frames contain some errors. Thus, we use an ensemble position-based error correcting algorithm to estimate the errors and correct them. The experimental results show that the proposed watermarking algorithm is imperceptible and moreover is robust against various attacks. After embedding 64 bits of watermark into each frame, the average peak signal-to-noise ratio between original frames and embedded frames is 45.7 dB.

• Journal of Broadcast Engineering
• /
• v.16 no.2
• /
• pp.247-257
• /
• 2011

In this paper, we propose a scene adaptive method to obtain two LDR images with proper shutter speeds which capture the irradiance of scene effectively. The proposed method adaptively selects two shutter speeds across the video frame even when the illumination varies continuously. For the performance evaluation, we compute the PNSR to the ground truth which is obtained by the state-of-the-art HDR imaging method. It shows that the proposed method is able to select approximately optimal shutter speeds while avoiding the exhaustive search of every possible pair of shutter speeds.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.1
• /
• pp.49-57
• /
• 2012

An image reconstruction algorithm is proposed to estimate transmitted original images from images displayed on a video terminal. The proposed algorithm acquires images that are displayed on video terminal screens by using a camera. The transmitted images are then estimated with the acquired images. However, camera-acquired images exhibit geometric and color distortions caused by characteristics of the camera and display devices. We make use of a geometric distortion correction algorithm that exploits homography and color distortions using a weighted-linear model. The experimental results show that the proposed algorithm yields promising estimation performance with respect to the peak signal-to-noise ratio (PSNR). PSNR values of the estimated images with respect to the corresponding original images range from 28-29 dB.

• ETRI Journal
• /
• v.26 no.5
• /
• pp.375-383
• /
• 2004

This paper proposes a common interleaving method associated with independent channel-encoding among transmitter antenna branches in orthogonal frequency and code division multiplexing based on multiple-input multiple-output (MIMO) multiplexing to achieve an extremely high throughput such as 1 Gbps using a 100 MHz bandwidth. This paper also investigates the average packet error rate performance as a function of the average received signal energy per bit-to-background noise power spectrum density ratio $(E_b/N_0)$. We found that the loss in the required average received $E_b/N_0$ of the proposed method is only within approximately 0.3 dB in up to a 12-path Rayleigh fading channel, using 16QAM and Turbo coding with a coding rate of 5/6. We also clarify that even for a large fading correlation among antenna branches, 1 Gbps is still possible by increasing the transmission power. Therefore, the proposed method reduces the processing rate to 1/4 in the turbo decoder with only a slight loss in the required average received $E_b/N_0$.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.2
• /
• pp.372-379
• /
• 2015

The latest video compression standard (such as H.264/AVC and HEVC) utilizes quarter-pel accuracy motion estimation in order to retain detailed motion information. Many sub-pixel motion estimation algorithms used in the spatial domain usually encounters increment of computational complexity due to embedded interpolation algorithm. In this paper, an approach to measure sub-pixel accuracy motion estimation in frequency domain using shifting matrix is proposed. Complexity can be reduced utilizing shifting matrix algorithm in frequency domain and simulation results demonstrate not only higher PSNR but lower bit rates than spatial domain algorithms.

• ETRI Journal
• /
• v.41 no.4
• /
• pp.415-425
• /
• 2019

Generative adversarial networks (GAN) have been successfully used in many image restoration tasks, including image denoising, super-resolution, and compression artifact reduction. By fully exploiting its characteristics, state-of-the-art image restoration techniques can be used to generate images with photorealistic details. However, there are many applications that require faithful rather than visually appealing image reconstruction, such as medical imaging, surveillance, and video coding. We found that previous GAN-training methods that used a loss function in the form of a weighted sum of fidelity and adversarial loss fails to reduce fidelity loss. This results in non-negligible degradation of the objective image quality, including peak signal-to-noise ratio. Our approach is to alternate between fidelity and adversarial loss in a way that the minimization of adversarial loss does not deteriorate the fidelity. Experimental results on compression-artifact reduction and super-resolution tasks show that the proposed method can perform faithful and photorealistic image restoration.

• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.2
• /
• pp.70.4-70.4
• /
• 2019

Image super-resolution (SR) is a technique that enhances the resolution of a low resolution image. In this study, we use three SR models (RCAN, ProSRGAN and Bicubic) for enhancing solar SDO/HMI magnetograms using deep learning. Each model generates a high resolution HMI image from a low resolution HMI image (4 by 4 binning). The pixel resolution of HMI is about 0.504 arcsec. Deep learning networks try to find the hidden equation between low resolution image and high resolution image from given input and the corresponding output image. In this study, we trained three models with HMI images in 2014 and test them with HMI images in 2015. We find that the RCAN model achieves higher quality results than the other two methods in view of both visual aspects and metrics: 31.40 peak signal-to-noise ratio(PSNR), Correlation Coefficient (0.96), Root mean square error (RMSE) is 0.004. This result is also much better than the conventional bi-cubic interpolation. We apply this model to a full-resolution SDO/HMI image and compare the generated image with the corresponding Hinode NFI magnetogram. As a result, we get a very high correlation (0.92) between the generated SR magnetogram and the Hinode one.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.11
• /
• pp.2103-2120
• /
• 2011

In this research, we consider the replication strategy for the applications of video streaming in wireless mesh networks (WMNs). In particular, we propose a closed-form of optimal replication densities for a set of frames of a video streaming by exploiting not only the skewed access probability of each frame but also the skewed loss probability and skewed encoding rate-distortion information. The simulation results demonstrate that our method improves the replication performance in terms of user-perceived quality (UPQ) which includes: 1) minimum average maximum reconstructed distortion for high peak signal-to-noise ratio (PSNR), 2) small reconstructed distortion fluctuation among frames for smooth playback, and 3) reasonable average maximum transmission distance for continuous playback. Furthermore, the proposed strategy consumes smaller storage capacity compared to other existing optimal replication strategies. More importantly, the effect of encoding rate is carefully investigated to show that high encoding rate does not always gain high performance of replication for video streaming.

• Journal of the Korea Computer Industry Society
• /
• v.4 no.12
• /
• pp.953-964
• /
• 2003

In the block matching algorithm, search patterns of different shapes or sizes and the distribution of motion vectors have a large impact on both the searching speed and the image quality. In this paper, we propose a new fast block matching algorithm using the cross pattern and the flat-hexagon search pattern. Our algorithm first finds the motion vectors that are close to the center of search window using the cross pattern, and then lastly finds the other motion vectors that are not close to the center of search window using the flat-hexagon search pattern. Through experiments, compared with the hexagon-based search algorithm(HEXBS), the proposed cross pattern and flat-hexagonal pattern search algorithm(CFHPS ) improves about 0.2-6.2% in terms of average number of search point per motion vector estimation and improves about 0.02-0.31dB in terms of PSNR(Peak Signal to Noise Ratio).

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.7
• /
• pp.2464-2479
• /
• 2014

Multi-view video coding is an international encoding standard that attains good performance by fully utilizing temporal and inter-view correlations. However, it suffers from high computational complexity. This paper presents a fast encoder design to reduce the level of complexity. First, when the temporal correlation of a group of pictures is sufficiently strong, macroblock-based inter-view prediction is not employed for the non-anchor pictures of B-views. Second, when the disparity between two adjacent views is above some threshold, frame-based inter-view prediction is disabled. Third, inter-view prediction is not performed on boundary macroblocks in the auxiliary views, because the references for these blocks may not exist in neighboring views. Fourth, finer partitions of inter-view prediction are cancelled for macroblocks in static image areas. Finally, when estimating the disparity of a macroblock, the search range is adjusted according to the mode size distribution of the neighboring view. Compared with reference software, these techniques produce an average time reduction of 83.65%, while the bit-rate increase and peak signal-to-noise ratio loss are less than 0.54% and 0.05dB, respectively.