• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.12
• /
• pp.4326-4344
• /
• 2021

The SinGAN is one of generative adversarial networks that can be trained on a single nature image. It has poor ability to learn more global features from nature image, and losses much local detail information when it generates arbitrary size image sample. To solve the problem, a non-linear function is firstly proposed to control downsampling ratio that is ratio between the size of current image and the size of next downsampled image, to increase the ratio with increase of the number of downsampling. This makes the low-resolution images obtained by downsampling have higher proportion in all downsampled images. The low-resolution images usually contain much global information. Therefore, it can help the model to learn more global feature information from downsampled images. Secondly, the attention mechanism is introduced to the generative network to increase the weight of effective image information. This can make the network learn more local details. Besides, in order to make the output image more natural, the TVLoss function is introduced to the loss function of SinGAN, to reduce the difference between adjacent pixels and smear phenomenon for the output image. A large number of experimental results show that our proposed model has better performance than other methods in generating random samples with fixed size and arbitrary size, image harmonization and editing.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.42 no.1
• /
• pp.46-49
• /
• 2017

In this letter, we investigate the impact of image downsampling on the performance of background subtraction in Full-HD soccer videos. To this end, we evaluated the performance of background subtraction in terms of both accuracy and computational time. Furthermore, for the sake of completeness, we used two different background subtraction methods under the same experimental setup. For the quantitative comparison, we employed the F-measure and FPS(frames per second). We believe that this study serves as a practically useful benchmark for researchers and practitioners in developing a fast background subtraction algorithm adopted for building real-time intelligent soccer video analysis systems.

• Journal of Broadcast Engineering
• /
• v.8 no.1
• /
• pp.30-36
• /
• 2003

According to an evolution of image and video compression technologies, most digital images are in the compressed form. Resizing of these compressed images have various applications such as transmission of resized image according to varying bandwidth, content adaptation for display purpose and etc. Discrete Cosine Transform (DCT) is the most popular transformation for image compression. Recently, several researches have been performed to obtain the reconstructed image of original size in the DCT domain after downsampling and upsampling in the DCT domain. Main focus of these researches is to improve quality of the reconstructed image after downsampling and upsampling in the DCT domain In this paper, we present an modified IDCT method to downsize DCT-encoded image. Furthermore, we propose an efficient scheme for image downsampling and upsampling in the DCT domain With these modified IDCT method. The proposed scheme Provides higher PSNR values than the existing schemes In terms of the reconstructed image after halving and doubling in the DCT domain.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.11a
• /
• pp.4-7
• /
• 2022

기존 합성곱 신경망 기반의 잡음 제거 네트워크들은 학습을 위한 noisy-clean 데이터 쌍을 필요로 한다. 하지만 실제 카메라 잡음의 경우, 잡음에 대한 깨끗한 원본 영상을 얻는 것은 불가능하거나 많은 비용이 소모된다. 따라서 이러한 방법을 해결하기 위하여 원본 영상 없이 잡음 영상만으로만 잡음 제거 네트워크를 학습하는 방법들이 제안되어왔다. 그 중 카메라 잡음 영상을 처리하기 위한 대표적인 방법으로 학습과 추론에서 비대칭적인 downsampling을 사용하는 AP-BSN이 제안되었다. 본 논문에서는 Functional neural network를 AP-BSN 알고리즘에 적용하여 다양한 downsampling ratio에 대응되는 하나의 네트워크를 학습하였다. 이를 통해 기존 hyperparameter로 사용되던 downsampling ratio에 대한 결과를 하나의 네트워크에서 분석 및 확인하였다. 또한 해당 파라미터를 조절함으로써 다양한 잡음 제거 후보들을 추출하고 사용자가 원하는 잡음 제거 정도를 조정할 수 있도록 하였다.

• Journal of KIISE:Software and Applications
• /
• v.30 no.7_8
• /
• pp.687-695
• /
• 2003

In this paper, we present a document image compression system based on segmentation, Quincunx downsampling, (5/3) wavelet lifting and subband-oriented SPIHT coding. We reduced the coding time by the adaptation of subband-oriented SPIHT coding and Quincunx downsampling. And to increase compression rate further, we applied arithmetic coding to the bitstream of SPIHT coding output. Finally, we present the reconstructed images for visual comparison and also present the compression rates and PSNR values under various scalar quantization methods.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.8
• /
• pp.1714-1719
• /
• 2011

High quality image processing schemes are used more widely than ever according to the development of various visual media. We need a compressed form of image for sending more capacity and a controlling strategy of images for small display devices. In this paper, we propose an image upsampling and downsamplig scheme using DCT coefficients for those purposes. Our scheme is designed to control the size of picture based on the target display media by reducing the data in DCT domain while not increasing the computational burdens. With the power of controlling the resolution in DCT domain, the proposed method shows higher PSNR than other competing methods in experiment.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.10C
• /
• pp.1484-1489
• /
• 2004

Given a video frame in terms of its 8${\times}$8 block-DCT coefncients, we wish to obtain a downsized or upsized version of this Dame also in terms of 8${\times}$8 block DCT coefficients. The DCT being a linear unitary transform is distributive over matrix multiplication. This fact has been used for downsampling video frames in the DCT domains in Dugad's, Mukherjee's, and Park's methods. The downsampling and upsampling schemes combined together preserve all the low-frequency DCT coefficients of the original image. This implies tremendous savings for coding the difference between the original frame (unsampled image) and its prediction (the upsampled image).This is desirable for many applications based on scalable encoding of video. In this paper, we extend the earlier works to various DCT sizes, when we downsample and then upsample of an image by a factor of two. Through experiment, we could improve the PSM values whenever we increase the DCT block size. However, because the complexity will be also increase, we can say there is a tradeoff. The experiment result would provide important data for developing fast algorithms of compressed-domain image/video resizing.

• Proceedings of the KIEE Conference
• /
• 2006.10c
• /
• pp.145-147
• /
• 2006

Image resizing is to change an image size by upsampling or downsampling of a digital image. Most still images and video frames are given in a compressed domain on digital media. Image resizing of a compressed image can be performed in a spatial domain via decompression or recompression. In general, resizing of a compressed image in a compressed domain is much faster than that in a spatial domain. In this paper, we propose an approach to resize images in the integer discrete cosine transform (DCT) domain, which exploits the multiplication-convolution property of DCT.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.7
• /
• pp.1624-1632
• /
• 1996

Wavelet transform is widely used for high compression ratio image compression. It requeires a large memory when it is implemented by a hardware. Therefore, it is efficient to divide the entire image into blocks. Because the wavelet transform for divided blocks causes losses, pixels of the adjacent blocks are used. In the case of color image compression, the image is decomposed into brightness and color components, and then color components are downsampled. When the wavelet transform is performed by using pixels of adjacentblocks, the number of necessary pixels are doubled due to downsampling of color components. In this paper, we propose an efficient block wavelet transform using variablefilter length for brightness and color components. By using the proposed method, the number of pixels of adjacent blocks is optimized. We show the degradation of image quality due to the reduction of filter length for color components is negligible through simulations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.1
• /
• pp.56-67
• /
• 2013

Because H.264/SVC can meet the needs of different networks and user terminals, it has become more and more popular. In this paper, we focus on the spatial resolution scalability of H.264/SVC and propose a blind video watermarking algorithm for the copyright protection of H.264/SVC coded video. The watermark embedding occurs before the H.264/SVC encoding, and only the original enhancement layer sequence is watermarked. However, because the watermark is embedded into the average matrix of each macro block, it can be detected in both the enhancement layer and base layer after downsampling, video encoding, and video decoding. The proposed algorithm is examined using JSVM, and experiment results show that is robust to H.264/SVC coding and has little influence on video quality.