• 전자공학회논문지CI
• /
• 제41권6호
• /
• pp.91-99
• /
• 2004

본 논문은 H.264와 MPEG-4 간의 원활한 통신을 하기위한 두 가지 트랜스코딩 방법을 제안한다. 같은 공간적 시간적 해상도(spatio-temporal resolution)를 유지하는 트랜스코팅 방법과 공간적 해상도(temporal resolution)를 줄이는 트랜스코팅 방법을 제안한다. H.264 비트스트림(bitstream)이 MPEG-4 비트스트림으로 변환 시 H.264 블록형태를 MPEG-4에서 사용 할 수 있는 블록형태로 변환 시켜야 하며, 4×4 블록단위의 움직임 벡터도 8×8 블록단위의 움직임 벡터로 조정하여야 한다. 두 가지 제안된 트랜스코딩 방법은 직렬 화소영역 트랜스코팅 방법(cascade pixel-domain transcoding) 보다 MPEG-4 부호화기 측에서 4.1~5.1배 부호화 속도가 빠를 뿐만 아니라 영상의 화질 저하는 최고 0.3dB정도 밖에 떨어 지지 않는다.

• IEIE Transactions on Smart Processing and Computing
• /
• 제5권2호
• /
• pp.129-142
• /
• 2016

In Wyner-Ziv coding, video signals are reconstructed by correcting side information generated by block-based motion estimation/compensation at the decoder. The correction is not always accurate due to the limited number of parity bits and early stopping of low-density parity check accumulate (LDPCA) decoding in distributed video coding, or due to the limited number of measurements in distributed compressive video sensing. The blocking artifacts caused by block-based processing are usually conspicuous in smooth areas and degrade the perceptual quality of the reconstructed video. Conventional deblocking filters try to remove the artifacts by treating both sides of the block boundary equally; however, coding errors generated by block-based processing are not necessarily the same on both sides of the block boundaries. Such a block-wise difference is exploited in this paper to improve deblocking for Wyner-Ziv frameworks by designing a filter where the deblocking strength at each block can be non-identical, depending on the reliability of the reconstructed pixels. Test results show that the proposed filter not only improves subjective quality by reducing the coding artifacts considerably, but also gains rate distortion performance.

• ETRI Journal
• /
• 제31권1호
• /
• pp.21-30
• /
• 2009

In low bit-rate video transmission, the payload of a single packet can often contain a whole coded frame due to the high compression ratio in both spatial and temporal domains of most modern video coders. Thus, the loss of a single packet not only causes the loss of a whole frame, but also produces error propagation into subsequent frames. In this paper, we propose a novel whole frame error concealment algorithm which reconstructs the first of the subsequent frames instead of the current lost frame to suppress the effects of error propagation. In the proposed algorithm, we impose a constraint which uses side match distortion (SMD) and overlapped region difference (ORD) to estimate motion vectors between the target reconstructed frame and its reference frame. SMD measures the spatial smoothness connection between a block and its neighboring blocks. ORD is defined as the difference between the correlated pixels which are predicted from one reference pixel. Experimental results show that the proposed algorithm effectively suppresses error propagation and significantly outperforms other conventional techniques in terms of both peak signal-to-noise ratio performance and subjective visual quality.

• 음성과학
• /
• 제15권2호
• /
• pp.67-77
• /
• 2008

In this paper, we implemented a robust speech recognizer using the TMS320VC33 DSP. For this implementation, we had built speech and noise database suitable for the recognizer using spectral subtraction method for noise removal. The recognizer has an explicit structure in aspect that a speech signal is enhanced through spectral subtraction before endpoints detection and feature extraction. This helps make the operation of the recognizer clear and build HMM models which give minimum model-mismatch. Since the recognizer was developed for the purpose of controlling car facilities and voice dialing, it has two recognition engines, speaker independent one for controlling car facilities and speaker dependent one for voice dialing. We adopted a conventional DTW algorithm for the latter and a continuous HMM for the former. Though various off-line recognition test, we made a selection of optimal conditions of several recognition parameters for a resource-limited embedded recognizer, which led to HMM models of the three mixtures per state. The car noise added speech database is enhanced using spectral subtraction before HMM parameter estimation for reducing model-mismatch caused by nonlinear distortion from spectral subtraction. The hardware module developed includes a microcontroller for host interface which processes the protocol between the DSP and a host.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2014년도 추계학술대회
• /
• pp.15-17
• /
• 2014

본 논문에서는 하드웨어의 제한된 자원을 이용하여 HEVC 코덱을 구현할 때 DCT 와 엔트로피 부호화를 사용하지 않고 율 및 왜곡값을 예측하여 고효율의 부호화를 수행하는 방법에 대하여 제안한다. HEVC 는 기존의 부호화기에 비하여 계층적 부호화 구조와 함께 큰 블록 크기를 갖는 DCT 와 엔트로피 부호화를 반복적으로 수행하기 때문에 하드웨어 구현 시 그 복잡도가 매우 크게 증가한다. 먼저 DCT 는 하다마드변환 행렬과 또 다른 정규 직교 변환 행렬의 곱으로 표현될 수 있는 성질을 이용하여 부호화 변환 시 생성된 하드마드변환 행렬에 저복잡도의 정규 직교 변환 행렬을 곱하여 DCT 변환 계수를 생성한 후 변환 및 양자화를 수행한다. 왜곡값의 경우, 이 때 생성된 양자화 계수와 변환 계수 간의 차이를 변환도메인에서 제곱합을 이용하여 계산하여 역변환을 생략함으로써 복잡도를 감소시킬 수 있다. 또한 텍스처에 대한 비트율 예측은 각 CU 블록내의 양자화 계수의 수를 더하여 계산하여 엔트로피를 수행하지 않고 예측할 수 있다. 그리고 비 텍스처에 대한 비트율 예측의 경우 움직임벡터의 비트에 대한 Pseudo CABAC 코드를 수행하여 예측할 수 있다. 이러한 저 복잡도의 텍스처 및 비텍스처 비트와 왜곡을 예측함으로써 하다마드변환만을 이용하여 부호화하였을 때에 비해 최대 33%의 비트율 감소를 얻을 수 있었다.

• 한국멀티미디어학회논문지
• /
• 제23권2호
• /
• pp.186-203
• /
• 2020

Image-based Virtual Try-On Systems are among the most potential solution for virtual fitting which tries on a target clothes into a model person image and thus have attracted considerable research efforts. In many cases, current solutions for those fails in achieving naturally looking virtual fitted image where a target clothes is transferred into the body area of a model person of any shape and pose while keeping clothes context like texture, text, logo without distortion and artifacts. In this paper, we propose a new improved image-based virtual try-on network system based on keypoints, which we name as KP-VTON. The proposed KP-VTON first detects keypoints in the target clothes and reliably predicts keypoints in the clothes of a model person image by utilizing a dense human pose estimation. Then, through TPS transformation calculated by utilizing the keypoints as control points, the warped target clothes image, which is matched into the body area for wearing the target clothes, is obtained. Finally, a new try-on module adopting Attention U-Net is applied to handle more detailed synthesis of virtual fitted image. Extensive experiments on a well-known dataset show that the proposed KP-VTON performs better the state-of-the-art virtual try-on systems.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.59-62
• /
• 2009

In general, as a dynamic range of digital still camera is narrower than a real scene‘s, it is hard to represent the shadow region of scene. Thus, multi-scaled retinex algorithm is used to improve detail and local contrast of the shadow region in an image by dividing the image by its local average images through Gaussian filtering. However, if the chromatic distribution of the original image is not uniform and dominated by a certain chromaticity, the chromaticity of the local average image depends on the dominant chromaticity of original image, thereby the colors of the resulting image are shifted to a complement color to the dominant chromaticity. In this paper, a modified multi-scaled retinex method to reduce the influence of the dominant chromaticity is proposed. In multi-scaled retinex process, the local average images obtained by Gaussian filtering are divided by the average chromaticity values of the original image in order to reduce the influence of dominant chromaticity. Next, the chromaticity of illuminant is estimated in highlight region and the local average images are corrected by the estimated chromaticity of illuminant. In experiment, results show that the proposed method improved the local contrast and detail without color distortion.

• 한국정밀공학회지
• /
• 제16권10호
• /
• pp.80-86
• /
• 1999

Since the quality of the coated wires is in various applications dependant on the coating depth, accuracy of hole size of dies used for coating wires must be maintained precisely, in general within one micron. This paper proposes a new vision system which measures automatically the size and shape of small holes having diameters less than 1mm within an error limit of 1 micron. To quickly obtain the focused image, this paper proposes an estimation method of the camera position using only a couple of defocused hole images. It measures the distributions of light intensity around the image boundary and decides the direction and distance of a camera motion. The proposed system measures the size, shape distortion, inclination of the hole against the axis of the dies structure, to decides the acceptability of the dies for use. The proposed algorithm has been implemented using a cheap 640${\times}$480 image system and has shown an average size error of 1micron when measuring the dieses having 0.1mm to 1.0mm diameters. It can be applied to the inspection of the size and position of holes in PCB, too.

• 한국전자파학회논문지
• /
• 제10권6호
• /
• pp.855-865
• /
• 1999

본 논푼에서는 주파수 선택적 레일리 페이딩 채널에서 16 QAM-TCM의 BER 성능을 개선하는 방법을 제시한다. 주파수 선택적 페이딩 환경에서는 ISI 등, 페이딩에 의한 성능열화가 극히 심하므로 이에 대한 보상 방안으로 가우시안 보상기법을 이용한 다중 파일롯 심별에 의한 보상 방법을 제안하였다. 또한 대역폭 확장 없이 우수한 부호화 이득을 얻을 수 있는 트렐리스 부호화 방법을 적용하였으며 연집 에러를 방지하고자 비트 리버설 블럭 인터리벙 기법을 개선하여 적용하였다. 컴퓨터 시율레이션 결과 주파수 선택적 레일리 페이 딩 환경에서 기존의 PSAM 방법을 적용한 것보다 우수한 성능 개선 효과를 확인 할 수 있었다

• 한국CDE학회논문집
• /
• 제12권4호
• /
• pp.255-262
• /
• 2007

In the finite element analysis of forming process, objects are described with a finite number of elements and nodes and the approximated solutions can be obtained by the variational principle. One of the shortcomings of a finite element analysis is that the structure of mesh has become inefficient and unusable because discretization error increases as deformation proceeds due to severe distortion of elements. If the state of current mesh satisfies a certain remeshing criterion, analysis is stopped instantly and resumed with a reconstructed mesh. In the study, a new remeshing algorithm using tetrahedral elements has been developed, which is adapted to the desired mesh density. In order to reduce the discretization error, desired mesh sizes in each lesion of the workpiece are calculated using the Zinkiewicz and Zhu's a-posteriori error estimation scheme. The pre-constructed mesh is constructed based on the modified point insertion technique which is adapted to the density function. The object domain is divided into uniformly-sized sub-domains and the numbers of nodes in each sub-domain are redistributed, respectively. After finishing the redistribution process of nodes, a tetrahedral mesh is reconstructed with the redistributed nodes, which is adapted to the density map and resulting in good mesh quality. A goodness and adaptability of the constructed mesh is verified with a testing measure. The proposed remeshing technique is applied to the finite element analyses of forging processes.