• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.16 no.2
• /
• pp.172-179
• /
• 1991

In this paper we have proposed an efficient coding algorithm using directional filtering. First an image is transformed by using DCT which has better energy compaction and then the transformed image is divided into a low frequency component and several high frequency components. The transformed coefficients of each parts are transmitted respectively by using huffman code and these are transformed inversely at receiver. For the directional components total edge images are reconstructed at zero crossing points. We are able to reduce the amount of data by getting of complex component and making directional angles 90. As a results, this proposed method is better than that of Kunt in respect of processing time and memories. We have 38dB of image quality with objective measurement of PSNR and 0.26bpp of compression ratio which is acceptable.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.14 no.2
• /
• pp.11-16
• /
• 2014

Block truncation coding (BTC) is a recent technique used for compression of monochrome image data. The original algorithm of BTC implement to use the standard mean and the standard deviation. For data hiding, it is needed to modify bit-planes of an images. These modifications yields unacceptable quality of BTC images. Thus, we propose IODBTC (Improved Ordered dithering BTC) improving BTC. In this paper, we improve this problem. Experimental results show that the reconstructed secret messages are the same as the original secret messages, and that the proposed scheme exhibits a good BTC images compared to that of previous schemes.

• The KIPS Transactions:PartB
• /
• v.8B no.2
• /
• pp.208-214
• /
• 2001

웨이블릿을 이용한 분해와 복원은 영상의 압축에 널리 사용되는 도구이다. 웨이블릿 분해된자연 영상은 원 영상의 동일 부분을 표현하는 대역간의 계수 값 사이에 높은 상관관계를 갖고 있고, 하나의 대역 안의 인접한 계수들 사이에도 높은 상관 관계를 갖고 있다. EZW(Evbedded Zerotree Wavelet)나 SPIHT(Set Partitioning in Hierachical Trees)과 같이 성능이 우수하고 널리 사용되는 웨이블릿 기반 영상 압축은 웨이블릿 계수의 대역간 상관 관계를 이용하고 있다. Zero Tree에 기초한 이와 같은 방법들은 계수값의 부모 후손 관계를 이용하기 때문에 대역별 부호화를 수행할 수 없으며, 다중 해상도 복원이 필요한 응용 분야에 효과적으로 적용되어지지 못한다. 본 논문에서는 하나의 대역 내에서 인접한 계수들 간의 상관성을 이용한 대역 독립적 부호화 기법을 제안한다. 압축 성능은 SPIHT과 대등하거나 약간 우수한 것으로 나타났고, 다중해상도 복원에 효과적으로 적용할 수 있음을 확인하였다.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.15 no.3
• /
• pp.1-6
• /
• 1978

This paper deals with speech digitization and bandwidth compression techniques, particularly two predictive coding methods-namely, adaptive diferentia1 pulse code modulation(ADPCM) and adaptive delta modulation (ADM). The principle of a typical adoptive quantizer that is used in ADPCM is explained, and two analysis methods for the adaptive predictor coefficents, block and sequential analyses, are discussed. Also, three companding methods (instantaneous, syllabic, and hybrid companding) that are used in ADM are explained in detail, and their performances are compared. In addition, the performances of ADPCM and ADM as speech coders are compared, and the merits of each coder are discussed.

• Proceedings of the IEEK Conference
• /
• 2000.11c
• /
• pp.97-100
• /
• 2000

The most important technology is the compression methods in the multimedia society. Audio files are rapidly propagated through internet. MP-3(MPEG-1 Layer3) is offered to CD tone quality in 128kbps, but 64kbps below tone-quality is abruptly down. On the other hand, MPEG-II AAC (Advanced Audio Coding) is not compatible with MPEG-I, but AAC has a high compression ratio 1.4 times better than MP-3 and it has max. 7.1 channel and 96KHz sampling rate. In this paper, we designed the optimized MDCT (Modified Discrete Cosine Transform) that could decrease the capacity of enormous computation and could increase the processing speed in the MPEG-2 AAC encoder.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.11 no.4
• /
• pp.235-241
• /
• 2016

Compared to the former H.264 standard, the number of the prediction modes has highly increased in HEVC intra prediction. Compression efficiency and accurate prediction are significantly improved. However, the computational complexity increases as well. To solve this problem, this paper proposes the new scheme where not only prediction modes but also block partition candidate are early chosen. Compared to the original intra prediction in HEVC, the proposed scheme achieves about 38% reduction in processing cycles with a marginal loss in compression efficiency.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.5
• /
• pp.1025-1030
• /
• 2005

With the increasing use of multimedia technologies, image compression requires higher performance as well as new functionality. Specially, in the specific area of still image encoding, a new standard, JPEG2000 was developed. This paper proposed Set Partitioning Sorting Algorithm that uses a method to optimized selection of threshold from feature of wavelet transform coefficients and to removes sign bit in LL area on JPEG2000. Experimental results show the proposed algorithm achieves more improved bit rate.

• Convergence Security Journal
• /
• v.7 no.2
• /
• pp.57-63
• /
• 2007

The security of data in network is provided by encryption. Selective encryption is a recent approach to reduce the computational cost and complexity for large file size data such as image and video. This paper describes techniques to encrypt Huffman code and discusses the performance of proposed scheme. We propose a simple encryption technique applicable to the Huffman code and study effectiveness of encryption against insecure channel. Our scheme combine encryption process and compression process, and it can reduce processing time for encryption and compression by combining two processes.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.7
• /
• pp.1613-1623
• /
• 1996

In this paper, an image coding scheme using removal of isolated coefficients and quadtree structure in wavelet transformed domain, is suggested. The number of significant coefficients can be reduced by quantizing in different stages according to the weights of the each band. The quadtree structure preseves the location information while significantly reducing the number of bits required to represent the locality of an image. To increase the efficienty of the wavelet transformed images using quadtree, the energy of the isolated coefficients is analyzed and those with low energy are eliminated. Furthermore, to control the bit rate the entropy of each block is sorted according to its energy concentration and the number of bits required for encoding, and the blocks with low energy are removed to achieve the desired compression rate.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.143-145
• /
• 2022

MPEG-VCM(Video Coding for Machine)은 입력된 비디오 특징(feature)를 압축하는 Track1 과 입력 영상을 직접 압축하는 Track2 로 나뉘어 표준화가 진행중이다. 본 논문은 VCM Track 1 에 해당하는 Detectron2 FPN(Feature Pyramid Network)에서 추출한 다중 스케일 특징맵을 VVC 로 압축하는 MSFC(Multi-Scale Feature Compression)을 구조를 제안한다. 본 논문의 MSFC 에서는 다중 스케일 특징을 결합하여 부호화/복호화하는 기존의 구조에서 특징맵의 해상도를 줄여 압축하는 개선된 MSFC 를 제시한다. 제안 방법은 VCM 의 Track2 의 영상 앵커(image anchor) 보다 우수한 BPP-mAP 성능을 보이고 최대 -84.98%의 BD-rate 성능향상을 보인다.