• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.2-4
• /
• 2021

In a distributed machine learning system, sharing gradients was considered safe because it did not share original training data. However, recent studies found that malicious attacker could completely restore the original training data from shared gradients. Gradient Leakage Attack is a technique that restoring original training data by exploiting theses vulnerability. In this study, we present the image transformation method based on Discrete Cosine Transform to defend against the Gradient Leakage Attack on the federated learning setting, which training in local devices and sharing gradients to the server. Experiment shows that our image transformation method cannot be completely restored the original data from Gradient Leakage Attack.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.9
• /
• pp.30-40
• /
• 2014

As the Hadamard transform can be generalized into the Jacket transform, in this paper, we generalize the Haar transform into the Jacket-Haar transform. The entries of the Jacket-Haar transform are 0 and ${\pm}2^k$. Compared with the original Haar transform, the basis of the Jacket-Haar transform is general and more suitable for signal processing. As an application, we present the DCT-II(discrete cosine transform-II) based on $2{\times}2$ Hadamard matrix and HWT(Haar Wavelete transform) based on $2{\times}2$ Haar matrix, analysis the performances of them and estimate them via the Lenna image simulation.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.7
• /
• pp.487-494
• /
• 2013

Medical image watermarking has received extensive attention as wide security services in the healthcare information system. This paper proposes a blind medical image watermarking approach on the segmented gray-matter (GM) images by utilizing discrete wavelet transform (DWT) and discrete cosine transform (DCT) along with enhanced suppressed fuzzy C-means (EnSFCM) for the optimal selection of sub-blocks position to insert a watermark. Experimental results show that the proposed approach outperforms other methods in terms of peak signal to noise ratio (PSNR) and M-SVD. In addition, the proposed approach shows better robustness than other methods in normalized correlation (NC) values against several attacks, such as noise addition, filtering, JPEG compression, blurring, histogram equalization, and cropping.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.3
• /
• pp.25-34
• /
• 2004

The ability to perform the same operations in the transform domain as in the spatial domain is important for efficient image transmission through a channel. We perform image resizing, which includes magnifying and reducing the size, in the discrete cosine transform(DCT) domain and the effects of the transform domain approach are analyzed in the corresponding spatial domain. Based on this analysis, the two resizing algorithms are proposed. The first one further compresses the images encoded by the compression standard by reducing the size before compression, and the other reduces the loss of information while maintaining the conventional compression rate. Because of its compatibility with standard codec, these algorithms can be easily embedded in JPEG and MPEG codecs, which are widely used for the purpose of image storage and transmission. Experimental results show a reduction of about half the bit size with similar image quality and about a 2- or 3-dB quality improvements in the similar compression rate.

• Journal of the Institute of Convergence Signal Processing
• /
• v.14 no.1
• /
• pp.13-20
• /
• 2013

A combined DWT and DCT based watermarking technique of low frequency watermarking with weighted correction is proposed. The DWT has excellent spatial localization, frequency spread and multi-resolution characteristics, which are similar to the theoretical models of the human visual system (HVS). The DCT based watermarking techniques offer compression while DWT based watermarking techniques offer scalability. These desirable properties are used in this combined watermarking technique. In the proposed method watermark bits are embedded in the low frequency band of each DCT block of selected DWT sub-band. The weighted correction is also used to improve the imperceptibility. The extracting procedure reverses the embedding operations without the reference of the original image. Compared with the similar approach by DCT based approach and DWT based approach, the experimental results show that the proposed algorithm apparently preserves superiori mage quality and robustness under various attacks such as JPEG compression, cropping, sharping, contrast adjustments and so on.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.11a
• /
• pp.39-42
• /
• 2016

Since a display device such as TV or signage is getting larger, the types of media is getting changed into wider view one such as UHD, panoramic and jigsaw-like media. Especially, panoramic and jigsaw-like media is realized by stitching video clips, which are captured by different camera or devices. In order to stich those video clips, it is required to find out 2D Adjacency Matrix, which tells spatial relationships among those video clips. Discrete Cosine Transform (DCT), which is used as a compression transform method, can convert the each frame of video source from the spatial domain (2D) into frequency domain. Based on the aforementioned compressed features, 2D adjacency Matrix of images could be found that we can efficiently make the spatial map of the images by using DCT. This paper proposes a new method of generating 2D adjacency matrix by using DCT for producing a panoramic and jigsaw-like media through various individual video clips.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.10 no.3
• /
• pp.130-134
• /
• 1985

In two-dimensional discrete cosine transform(DCT) coding, the measurements of the distributions of the transform coefficients are important because a better approximation yields a smaller mean square distorition. This paper presents the results of distribution tests which indicate that the statistics of the AC coefficients are well approximated to a generalized Gaussian distribution whose shape parameter is 0.6. Furthermore, from a simulation of the DCT coding, it was shown that the above approximation yields a higher experimental SNR and a better agreement between theory and simulation than the Gaussian or Laplacian assumptions.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.39 no.4
• /
• pp.423-430
• /
• 2002

When JPEG, a standard of stopped image compression, is high compressed, the image is severely blocked. Since JPEG performs compression after taking DCT(Discrete Cosine Transform). It has a defect that the quality of image becomes low with aliasing in the case of high compression. Though transformation cipher method can have high compression rate, flame nay happen to quality of image by transformation and reverse transformation. In this paper, we use wavelet transform and fractal theory in order to solve these problems. After we apply these two methods to stopped image, we can get some good results, improvement of speed and compression rate, and elimination of blocking appearance. Besides, we show quality of restoration image is better than established one.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.4
• /
• pp.106-114
• /
• 2016

In this paper, we propose an efficient integer inverse transform structure for vp9 decoder. The proposed structure is a hardware structure which is easy to control and requires less hardware resources, and shares algorithms for realizing entire DCT(Discrete Cosine Transform), ADST(Asymmetric Discrete Sine Transform) and WHT(Walsh-Hadamard Transform) in vp9. The integer inverse transform for vp9 google model has a fast structure, named butterfly structure. The integer inverse transform for google C model, unlike universal fast structure, takes a constant rounding shift operator on each stage and includes an asymmetrical sine transform structure. Thus, the proposed structure approximates matrix coefficient values for all transform mode and is used to matrix operation method. With the proposed structure, shared operations for all inverse transform algorithm modes can be possible with reduced number of multipliers compared to the butterfly structure, which in turn manages the hardware resources more efficiently.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1717-1720
• /
• 2004

In this paper we propose a new scheme of watermarking using the discrete wavelet transform, the discrete cosine transform, and the performance evaluation function, which does not deteriorate image quality and have robustness to attacks such as compression and scaling. moreover even if a detected watermark, which is a bit sequence in this paper, has some error bits, it can be correctly recovered using correlation-based determination scheme.