• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.05a
• /
• pp.211-213
• /
• 2010

본 논문에서는 형태학적 처리 방법과 8 방향 윤곽선 추적을 이용하여 손금을 추출하는 방법을 제안한다. YCbCr컬러 공간에서 Y:65~255, Cb:25~255, Cr:130~255에 해당되는 피부색 임계치를 이용하여 손 영역을 추출한다. 추출된 손 영역에서 내부 픽셀의 3:1 이상, 전체 영상의 2:1이상인 손의 형태학적 정보와 8 방향 윤곽선 추적 기법을 이용하여 잡음을 제거한다. 잡음이 제거된 손 영상에서 스트레칭 기법과 소벨 마스크를 이용하여 에지를 추출한다. 추출된 에지 영상에서 블록 이진화 기법을 이용하여 이진화한 후에 가로와 세로가 각각 10픽셀 이상이고 20픽셀 이하인 손금의 형태학적 정보를 이용하여 잡음 및 손의 윤곽선을 제외한 손금을 추출한다. 추출된 손금에서 Labeling 기법을 이용하여 개별 손금의 중요선을 추출한다. 핸드폰 카메라에서 획득한 손바닥 영상을 대상으로 실험한 결과, 제안된 방법이 손금 추출에 효율적인 것을 확인할 수 있었다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.2
• /
• pp.127-132
• /
• 2019

In this paper, we propose a visual inspection method of automotive parts for injection molding to improve the appearance quality and productivity of automotive parts. Exterior inspection of existing injection molding automobile parts was generally done by manual sampling inspection by human. First, we applied the edge-tolerance vision inspection algorithm ([1] - [4]) for vision inspection of electronic components (TFT-LCD and PCB) And we propose a new visual inspection method to overcome the problem. In the proposed visual inspection, the inspection images of the parts to be inspected are aligned on the basis of the reference image of good quality. Then, after partial adaptive binarization, the binary block matching algorithm is used to compare the good binary image and the test binary image. We verified the effectiveness of the edge-tolerance vision check algorithm and the proposed appearance vision test method through various comparative experiments using actual developed equipment.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.11a
• /
• pp.285-289
• /
• 2009

본 논문에서는 현재 비디오 코덱 표준인 H.264/AVC 에서 사용하고 있는 정수형 변환을 두 번 사용하는 다차원 변환 즉, MDT(Multi-Dimensional Transform) 을 제안한다. 제안하는 알고리듬은 H.264/AVC 7가지 인터 모드중에서 $16{\times}16$, $8{\times}16$, $16{\times}8$블록 모드에 적용된다. H.264/AVC에서 사용하는 정수형 변환을 $4{\times}4$블록 단위로 적용하고 인접하는 4개의 $4{\times}4$블록으로부터 같은 주파수 위치의 계수값을 모아서 16개의 $4{\times}1$벡터 변환을 추가로 수행하는 방법이다. MDT를 수행함으로써 H.264/AVC 표준에서 사용되는 변환이 가지는 예측오차를 유지하면서 공간적 중복도를 추가로 줄일 수 있다. MDT 된 계수들을 정수형으로 양자화하고 양자화된 계수값을 효율적으로 주사하는 방법을 제안하여 압축효율을 높였다. 제안된 알고리듬의 제안하는 MDT 기반 방법은 H.264/AVC High profile 표준보다 평균 3.00%의 비트 절감을 보였다.

• The Transactions of the Korea Information Processing Society
• /
• v.2 no.6
• /
• pp.937-946
• /
• 1995

We proposed and efficient block segmentation and classification method for the document analysis using SGLDM(spatial gray level dependence matrix) and BP (back Propagation) neural network. Seven texture features are extracted directly from the SGLDM of each gray-level block image, and by using the nonlinear classifier of neural network BP, we can classify document blocks into 9 categories. The proposed method classifies the equation block, the table block and the flow chart block, which are mostly composed of the characters, out of the blocks that are conventionally classified as non-character blocks. By applying Sobel operator on the gray-level document image beforebinarization, we can reduce the effect of the background noises, and by using the additional horizontal-vertical smoothing as well as the vertical-horizontal smoothing of images, we can obtain an effective block segmentation that does not lead to the segmentation into small pieces. The result of experiment shows that a document can be segmented and classified into the character blocks of large fonts, small fonts, the character recognigible candidates of tables, flow charts, equations, and the non-character blocks of photos, figures, and graphs.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.8C
• /
• pp.1159-1174
• /
• 2004

In this paper, we present an efficient algorithm for region analysis of business card images acquired in a PDA by using DCT and information pixel density. The proposed method consists of three parts: region segmentation, information region classification, and text region classification. In the region segmentation, an input business card image is partitioned into 8 f8 blocks and the blocks are classified into information and background blocks using the normalized DCT energy in their low frequency bands. The input image is then segmented into information and background regions by region labeling on the classified blocks. In the information region classification, each information region is classified into picture region or text region by using a ratio of the DCT energy of horizontal and vertical edge components to that in low frequency band and a density of information pixels, that are black pixels in its binarized region. In the text region classification, each text region is classified into large character region or small character region by using the density of information pixels and an averaged horizontal and vertical run-lengths of information pixels. Experimental results show that the proposed method yields good performance of region segmentation, information region classification, and text region classification for test images of several types of business cards acquired by a PDA under various surrounding conditions. In addition, the error rates of the proposed region segmentation are about 2.2-10.1% lower than those of the conventional region segmentation methods. It is also shown that the error rates of the proposed information region classification is about 1.7% lower than that of the conventional information region classification method.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.1
• /
• pp.38-46
• /
• 2020

LOw Frequency Analysis Recording (LOFAR) and Demodulation of Envelop Modulation On Noise (DEMON) grams are bearing-time-frequency plots of underwater acoustic signals, to visualize features for passive sonar. Those grams are characterized by tonal components, for which conventional data coding methods are not suitable. In this work, a novel LOFAR/DEMON gram compression algorithm based on binary map and prediction methods is proposed. We first generate a binary map, from which prediction for each frequency bin is determined, and then divide a frame into several macro blocks. For each macro block, we apply intra and inter prediction modes and compute residuals. Then, we perform the prediction of available bins in the binary map and quantize residuals for entropy coding. By transmitting the binary map and prediction modes, the decoder can reconstructs grams using the same process. Simulation results show that the proposed algorithm provides significantly better compression performance on LOFAR and DEMON grams than conventional data coding methods.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.38 no.5
• /
• pp.474-485
• /
• 2001

This paper proposes a user-key-based fragile watermarking for detecting image modification. The embedding data in a form of binary image for authentication are inserted to the DCT coefficients of each block of the given image. To minimize possible exposure of being watermarked and the location of insertion, it is proposed to utilize a user-specific key in randomizing those information. Each DCT block hides one bit of data, all of which represent the user-specific authentication data. Experiments with 5 real images demonstrate that the proposed method not only detects whether there is modification or not, but also the actual location of modification with minimal visual deterioration. However, the proposed method has room for improvement against its loss of watermark by an attack of compression by more than 50%.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.6
• /
• pp.88-94
• /
• 2009

This paper proposes real time object detection and tracking algorithm that can be applied to security and supervisory system field. A proposed system is devide into object detection phase and object tracking phase. In object detection, we suggest Adaptive background subtraction method and Adaptive block based model which are advanced motion detecting methods to detect exact object motions. In object tracking, we design a multiple vehicle tracking system based on Kalman filtering. As a result of experiment, motion of moving object can be estimated. the result of tracking multipul object was not lost and object was tracked correctly. Also, we obtained improved result from long range detection and tracking.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.4
• /
• pp.714-725
• /
• 1997

This paper proposes an image compression method using the wavelet transform and bit-plane coding of wavelet coefficients. The hierarchical application of wavelet transform to an image produces one low resoluation(the subband with lowest frequency) image and several high frequency subbands. In the proposed method, the low resolution image is compressed by a lossless method at 8 bits per each coefficient. However, the high frequency subbands are decomposed into 8 bit planes. With an adptive block coding method, the decomposed bit planes are effectively compressed using localized edge information in each bit plane. In addition, the propsoed method can control bit rates by selectively eliminating lessimportant subbands of low significant bit planes. Experimental results show that the proposed scheme has better performance in the peak signal to noise ratio (PSNR) and compression rate than conventional image coding methods using the wavelet transform and vector quantization.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.48 no.2
• /
• pp.41-52
• /
• 2011

It is important to reduce computational complexity of de-blocking filter for real-time implementation, because it accounts for a great part of total computational complexity of the decoder. Because there are a lot of conditional branches and memory accesses in a decoding loop, it is not easy to speed up the de-blocking filter. Therefore, this paper presents a new algorithm of de-blocking filter minimizing conditional branches and memory accesses. The proposed structure of de-blocking filter enables filter operation to parallelize by software pipelining. The proposed optimization method was implemented on a TMS320DM6467 EVM board and we achieved approximately 46% cycle reduction, compared with that of FFmpeg.