• Journal of Broadcast Engineering
• /
• v.20 no.3
• /
• pp.408-413
• /
• 2015

Recently, According to increasing interest to original sound Karaoke instrument, MIDI type karaoke manufacturer attempt to make more cheap method instead of original recoding method. The specific method is to make the original sound accompaniment to remove only the voice of the singer in the singer music album. In this paper, a system to separate vocal components from music accompaniment for stereo recordings were proposed. Proposed system consists of two stages. The first stage is a vocal detection. This stage classifies an input into vocal and non vocal portions by using SVM with MFCC. In the second stage, selective frequency subtractions were performed at each frequency bin in vocal portions. In this case, it is determined in consideration not only the energies for each frequency bin but also the phase of the each frequency bin at each channel signal. Listening test with removed vocal music from proposed system show relatively high satisfactory level.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.6
• /
• pp.1-7
• /
• 2010

This paper presents a novel implementation on multimedia fingerprinting algorithm based on BCH (Bose-Chaudhuri-Hocquenghem) code. The evaluation is put in force the colluder detection to n-1. In the proposed algorit hm, the used collusion attacks adopt logical combinations (AND, OR and XOR) and average computing (Averaging). The fingerprinting code is generated as below step: 1. BIBD {7,4,1} code is generated with incidence matrix. 2. A new encoding method namely combines BIBD code with BCH code, these 2 kind codes are to be fingerprinting code by BCH encoding process. 3. The generated code in step 2, which would be fingerprinting code, that characteristic is similar BCH {15,7} code. 4. With the fingerprinting code in step 3, the collusion codebook is constructed for the colluder detection. Through an experiment, it confirmed that the ratio of colluder detection is 86.6% for AND collusion, 32.8% for OR collusion, 0% for XOR collusion and 66.4% for Averaging collusion respectively. And also, XOR collusion could not detect entirely colluder and on the other hand, AND and Averaging collusion could detect n-1 colluders and OR collusion could detect k colluders.

• Journal of KIISE:Software and Applications
• /
• v.29 no.1_2
• /
• pp.114-125
• /
• 2002

In this paper, we are proposing a hierarchical segmentation method that first segments the video data into units of shots by detecting cut and dissolve, and then decides types of camera operations or object movements in each shot. In our previous work[1], each picture group is divided into one of the three detailed categories, Shot(in case of scene change), Move(in case of camera operation or object movement) and Static(in case of almost no change between images), by analysing DC(Direct Current) component of I(Intra) frame. In this process, we have designed two-stage hierarchical neural network with inputs of various multiple features combined. Then, the system detects the accurate shot position, types of camera operations or object movements by searching P(Predicted), B(Bi-directional) frames of the current picture group selectively and hierarchically. Also, the statistical distributions of macro block types in P or B frames are used for the accurate detection of cut position, and another neural network with inputs of macro block types and motion vectors method can reduce the processing time by using only DC coefficients of I frames without decoding and by searching P, B frames selectively and hierarchically. The proposed method classified the picture groups in the accuracy of 93.9-100.0% and the cuts in the accuracy of 96.1-100.0% with three different together is used to detect dissolve, types of camera operations and object movements. The proposed types of video data. Also, it classified the types of camera movements or object movements in the accuracy of 90.13% and 89.28% with two different types of video data.

• The Journal of the Acoustical Society of Korea
• /
• v.32 no.1
• /
• pp.71-78
• /
• 2013

This paper describes a multiple audio watermarking using Quantization Index Modulation (QIM) on frequency phase and magnitude response. Proposed embedding procedure is composed of two stage. At the first stage, the watermark is embedded on the frequency phase response using QIM. In the second stage, the watermark is embedded using adaptive QIM with the step-size that is adaptively determined using the maximum value of the frequency magnitude response of every frame. The watermark is extracted by calculating the Euclidean distance as the blind detection. The proposed method is robust against most of attacks of audio watermark benchmarking. For the Fourier attacks, the proposed method shows over 95% recovery rate.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.5 no.1
• /
• pp.13-19
• /
• 2012

In this paper, we propose the pulmonary nodule detection method based on hierarchical 3D block analysis. The proposed system consists of two main part. In the first part, we select the block which is need to analysis. In the second part, we analysis the selected blocks. We extract the shape based features of the object in the selected blocks. Support Vector Machine is applied to the extracted features to classify into nodules and non-nodules.

• Journal of Broadcast Engineering
• /
• v.13 no.2
• /
• pp.236-244
• /
• 2008

We propose a new building detection and description algorithm for Lidar data and photogrammetric imagery using color segmentation, line segments matching, perceptual grouping. Our algorithm consists of two steps. In the first step, from the initial building regions extracted from Lidar data and the color segmentation results from the photogrammetric imagery, we extract coarse building boundaries based on the Lidar results with split and merge technique from aerial imagery. In the secondstep, we extract precise building boundaries based on coarse building boundaries and edges from aerial imagery using line segments matching and perceptual grouping. The contribution of this algorithm is that color information in photogrammetric imagery is used to complement collapsed building boundaries obtained by Lidar. Moreover, linearity of the edges and construction of closed roof form are used to reflect the characteristic of man-made object. Experimental results on multisensor data demonstrate that the proposed algorithm produces more accurate and reliable results than Lidar sensor.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.10b
• /
• pp.499-504
• /
• 2006

터널 내 화재 발생 시 대규모의 인명, 재산 피해가 발생하는데 이러한 상황을 조기에 탐지함으로써 피해를 최소화하기 위한 시스템이 필요하다. 또한 터널 내 설치된 CCTV를 사람이 24시간 감시하기에는 너무 어려운 점이 많다. 이에 따라 적절한 영상 처리를 통한 화염 및 연기 검출 시스템을 통해 경보를 알려줄 경우, 보다 편리하고 사람이 모니터 앞에 없을 때 화재 발생 시 화재를 검출할 수 있어 피해를 최소화 할 수 있다. 본 논문에서는 영상처리 기법을 이용하여 터널 안에서 발생한 화재 및 연기를 고속으로 탐지하기 위한 알고리즘을 제안하였다. 터널 안에서의 화재 탐지는 차량 조명 및 터널내의 조명등과 같은 여러 가지 상황에 의해 산불 탐지 알고리즘과 다른 독자적인 알고리즘의 개발이 요구된다. 본 논문에서 제시한 두 가지 알고리즘은 기존 알고리즘보다 정확한 위치 탐지와 초기 단계에서의 탐지가 가능하도록 되었다. 또한 우리는 실험 결과를 통해 각각의 성능을 비교함으로써 제시한 알고리즘의 타당성을 보여주었다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2004.05a
• /
• pp.215-218
• /
• 2004

본 논문은 실내외에서 얻어진 자연 영상으로부터 기울어진 영상을 바로 보정하고 텍스트를 추출하는 방법을 제안한다. 본 연구는 명도 이미지를 대상으로 모든 과정이 4단계로 이루어진다. 첫째 자연 영상에서 에지 검출 처리를 위한 전처리 및 Canny 에지 추출을 수행하며, 둘째 영상의 기울기를 추출하기 위해 허프변환에 대한 전처리와 후처리를 한 후, 셋째로 잡음영상과 선을 제거하고 텍스트 특징을 이용한 후보영역 검출을 한다. 마지막으로 텍스트 후보영역 안에서 지역적 이진화를 수행하여 불필요한 비텍스트 연결 요소를 추려내기 위해 두 가지 텍스트 추출 방법을 수행한다. 본 연구는 게시판, 교통표지판, 책 표지 등 100장의 자연영상을 대상으로 실험한 결과 텍스트 추출에서 90.3% 추출 정확도를 가졌으며, 기울어진 각도 추출에서도 94.3%의 높은 추출률을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.11a
• /
• pp.732-735
• /
• 2010

본 연구에서는 음악 데이터 베이스에서 노래의 특정 구간과 가장 유사한 구간을 검색하는 시스템을 제안한다. 제안된 시스템에서는 음악을 다차원 시계열 데이터로 간주하고, 음악의 조성 차이 및 템포(tempo) 차이를 고려한 음악의 유사도 계산 방법을 사용한다. 유사도 계산의 전처리 단계에서 조성 차이를 보정하고, 비트(beat)를 검출하며, 추출된 크로마그램(chromagram)을 검출된 비트와 동기화 하여 평균한다. 이후, 동적 시간 왜곡(DTW; dynamic time warping)을 사용하여 두 구간사이의 유사도를 계산한 후 계산된 유사도 순서로 정렬된 검색 결과를 출력한다. 사용자는 제안된 시스템을 사용하여 선택 구간 유사도 검색과 자동 유사 검색 결과로 도출된 구간 쌍을 검토하여 유사 구간을 보다 쉽게 찾을 수 있다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.8A
• /
• pp.786-795
• /
• 2002

In this paper, we present a histogram and moment-based vidoe scencd change detection technique using hierarchical Hidden Markov Models(HMMs). The proposed method extracts histograms from a low-frequency subband and moments of edge components from high-frequency subbands of wavelet transformed images. Then each HMM is trained by using histogram difference and directional moment difference, respectively, extracted from manually labeled video. The video segmentation process consists of two steps. A histogram-based HMM is first used to segment the input video sequence into three categories: shot, cut, gradual scene changes. In the second stage, a moment-based HMM is used to further segment the gradual changes into a fade and a dissolve. The experimental results show that the proposed technique is more effective in partitioning video frames than the previous threshold-based methods.