• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.9
• /
• pp.106-113
• /
• 2015

This paper integrates human visual system (HVS) characteristics into compressed sensing recovery of color images. The proposed visual weighting of each color channel in group-sparsity minimization not only pursues sparsity level of image but also reflects HVS characteristics well. Additionally, an edge-preserving filter is embedded in the scheme to remove noise while preserving edges of image so that quality of reconstructed image is further enhanced. Experimental results show that the average PSNR of the proposed method is 0.56 ~ 4dB higher than that of the state-of-the art group-sparsity minimization method. These results prove the excellence of the proposed method in both terms of objective and subjective qualities.

• Proceedings of the IEEK Conference
• /
• 2000.11d
• /
• pp.175-178
• /
• 2000

The object level image compression is a useful technology for reducing the necessary data and manipulating individual objects. In this paper, we propose a new image object compression algorithm that uses the quadratic programming (QP) method to reduce the compressed data. The results indicate the superiority of the proposed QP based algorithm over the low pass extrapolation (LPE) method of MPEG-4.

• 한국정보통신설비학회:학술대회논문집
• /
• 2008.08a
• /
• pp.193-196
• /
• 2008

In this paper a detail study of existing steganographic methods are presented. An example is given of LSB substitution with uncompressed domain i.e., BMP image. In case of compressed domain JPEG image steganography is presented. Almost all popular steganographic algorithms, such as JPEG JSteg, F3, F4 and Selective Block Steganography (SBS) are described. The applications of steganographic methods are also presented briefly.

• Journal of Korea Multimedia Society
• /
• v.23 no.8
• /
• pp.1019-1029
• /
• 2020

In this paper, we propose a novel knowledge distillation algorithm to create an compressed deep ensemble network coupled with the combined use of local and global features of face images. In order to transfer the capability of high-level recognition performances of the ensemble deep networks to a single deep network, the probability for class prediction, which is the softmax output of the ensemble network, is used as soft target for training a single deep network. By applying the knowledge distillation algorithm, the local feature informations obtained by training the deep ensemble network using facial subregions of the face image as input are transmitted to a single deep network to create a so-called compressed ensemble DCNN. The experimental results demonstrate that our proposed compressed ensemble deep network can maintain the recognition performance of the complex ensemble deep networks and is superior to the recognition performance of a single deep network. In addition, our proposed method can significantly reduce the storage(memory) space and execution time, compared to the conventional ensemble deep networks developed for face recognition.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.01a
• /
• pp.520-524
• /
• 2009

In compressed color images, colors are usually represented by luminance and chrominance (YCbCr) components. Considering characteristics of human vision system, chrominance (CbCr) components are generally represented more coarsely than luminance component. Aiming at possible recovery of chrominance components, we propose a model-based chrominance estimation algorithm where color images are modeled by a Markov random field (MRF). A simple MRF model is here used whose local conditional probability density function (pdf) for a color vector of a pixel is a Gaussian pdf depending on color vectors of its neighboring pixels. Chrominance components of a pixel are estimated by maximizing the conditional pdf given its luminance component and its neighboring color vectors. Experimental results show that the proposed chrominance estimation algorithm is effective for quality improvement of compressed color images such as JPEG and JPEG2000.

• Journal of Institute of Control, Robotics and Systems
• /
• v.12 no.6
• /
• pp.585-593
• /
• 2006

In this paper, we propose a vision-based simultaneous localization and map-building (SLAM) algorithm. SLAM problem asks the location of mobile robot in the unknown environments. Therefore, this problem is one of the most important processes of mobile robots in the outdoor operation. To solve this problem, Extended Kalman filter (EKF) is widely used. However, this filter requires computational power (${\sim}O(N)$, N is the dimension of state vector). To reduce the computational complexity, we applied compressed extended Kalman filter (CEKF) to stereo image sequence. Moreover, because the mobile robots operate in the outdoor environments, we should estimate full d.o.f.s of mobile robot. To evaluate proposed SLAM algorithm, we performed the outdoor experiments. The experiment was performed by using new wheeled type mobile robot, Robhaz-6W. The performance results of CEKF SLAM are presented.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.97-99
• /
• 2022

Recently, the amount of data is increasing with the development of industries and technologies. Research on the processing and transmission of large amounts of data is attracting attention. Therefore, in this paper, compressed sensing was used to reduce the amount of data and its effect on Mask R-CNN algorithm was analyzed. We confirmed that as the compressed sensing rate increases, the amount of data in the image and the resolution decreases. However, it was confirmed that there was no significant degradation in the performance of object detection.

• Proceedings of the Korean Printing Society Conference
• /
• 2000.04a
• /
• pp.17-24
• /
• 2000

The Joint Photographic Experts Group (JPEG) is a standara still-image compression technique, established by the International for Standardization (ISO) and International Telecommunication Standardization Sector (ITUT). The standard is intended to be utilized in the various kinds of color still imaging systems as a standard color image coding format. Because JPEG is a lossy compression, the decompressed image pixel values are nto the same as values before compression. Image of JPEG compression is often made to JPEG recompression at saving to apply JPEG compression of color image. In general, JPEG is a lossy compression and compression image is predicted to be varied image quality according to recompressed Q-factor. Various distortions of JPEG compression and JPEG recompression has been reported in previous paper. In this paper, we compress four difference color samples (photo image, gradient image, vector drawing image, text image) according to various Q-factor, and then compressed images are recompressed according to various Q-factor once again. As the results, we inspect variation of quality and file size of recompressed color image, and ensure the optimum recompression factor.

• Smart Structures and Systems
• /
• v.25 no.3
• /
• pp.369-384
• /
• 2020

A wavefield sparse reconstruction technique based on compressed sensing is developed in this work to dramatically reduce the number of measurements. Firstly, a severely underdetermined representation of guided wavefield at a snapshot is established in the spatial domain. Secondly, an optimal compressed sensing model of guided wavefield sparse reconstruction is established based on l₁-norm penalty, where a suite of discrete cosine functions is selected as the dictionary to promote the sparsity. The regular, random and jittered undersampling schemes are compared and selected as the undersampling matrix of compressed sensing. Thirdly, a gradient projection method is employed to solve the compressed sensing model of wavefield sparse reconstruction from highly incomplete measurements. Finally, experiments with different excitation frequencies are conducted on an aluminum plate to verify the effectiveness of the proposed sparse reconstruction method, where a scanning laser Doppler vibrometer as the true benchmark is used to measure the original wavefield in a given inspection region. Experiments demonstrate that the missing wavefield data can be accurately reconstructed from less than 12% of the original measurements; The reconstruction accuracy of the jittered undersampling scheme is slightly higher than that of the random undersampling scheme in high probability, but the regular undersampling scheme fails to reconstruct the wavefield image; A quantified mapping relationship between the sparsity ratio and the recovery error over a special interval is established with respect to statistical modeling and analysis.

• The KIPS Transactions:PartC
• /
• v.8C no.5
• /
• pp.525-534
• /
• 2001

Generally, compression and encryption procedures are performed independently in lossless image compression and encryption. When compression is followed by encryption, the compressed-stream should have the property of randomness because its entropy is decreased during the compression. However, when full data is compressed using image compression methods and then encrypted by encryption algorithms, real-time processing is unrealistic due to the time delay involved. In this paper, we propose to combine compression and encryption to reduce the overall processing time. It is method decomposing gray-scale image by means of quadtree compression algorithms and encrypting the structural part. Moreover, the lossless compression ratio can be increased using a transform that provides an decorrelated image and homogeneous region, and the encryption security can be improved using a reconstruction of the unencrypted quadtree data at each level. We confirmed the increased compression ratio, improved encryption security, and real-time processing by using computer simulations.