• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2005년도 춘계 학술대회 발표논문집
• /
• pp.49-52
• /
• 2005

Cepstral Mean Subtraction (CMS) makes effectively compensation for a channel distortion, but there are some shortcomings such as distortions of feature parameters, waiting for the whole speech sentence. By assuming that the silence parts have the channel characteristics, we consider the channel normalization using subtraction of cepstral means which are only obtained in the silence areas. If the considered techniques are successfully used for the channel compensation, the proposed method can be used for real time processing environments or time important areas. In the experiment result, however, the performance of our method is not good as CMS technique. From the analysis of the results, we found potentiality of the proposed method and will try to find the technique reducing the gap between CMS and ours method.

• 한국시뮬레이션학회논문지
• /
• 제6권1호
• /
• pp.53-60
• /
• 1997

Results from a nonequilibrium molecular dynamics (NEMD) simulation are presented for an argon liquid subject to a shear flow. The segmented molecular dynamics method and the subtraction technique used in NEMD program to reduce the thermal fluctuation noise in data are studied with different shear rates. The standard deviation in the shear stress reduced from 0.030 to 0.004 by the segmented molecular dynamics method for 50 repeated segments. On the other hand, the standard deviation of the data remained the same when the subtraction technique was applied, where as the results of shear stress by constant value in a random way.

• 치과방사선
• /
• 제17권1호
• /
• pp.175-181
• /
• 1987

The purpose of this study is to know the value of the photographic subtraction technic in the transcranial oblique lateral projection of the TMJ. The author examined the transcranial oblique lateral projection radiographs which comprise 50 cases of 32 persons, compared the transcranial oblique lateral projection films and those subtraction films. The following results were obtained. 1) The condyle at closed jaw position had showed a reversed tone image, but the condyle at the opening position had showed a re-reversed ordinary image. Both condyles had showed one subtraction film because radiographic interpretation of TMJ was easy. 2) On 50 cases of subtraction films, 46 cases had showed same radiographic images compared with trascranial oblique lateral projection films. Four cases (3 cases of erosion, 1 cases of sclerosis) had showed additional changes of images so that capability of interpretation was improved.

• 한국음향학회지
• /
• 제18권1호
• /
• pp.38-44
• /
• 1999

본 연구에서는 단일 채널 단구간 진폭 스펙트럼 추정 기법의 하나인 Spectral Subtraction 방법과 2 채널 Griffiths-Jim Beamformer를 결합한 음성 강조기법을 제안한다. 기존의 단구간 진폭 스펙트럼 추정 기법에서는 관측된 신호의 스펙트럼에서 잡음의 평균 스펙트럼을 감산하여 잡음을 제거하고 있지만, 이 방법을 이용하여 잡음을 제거 할 경우에는 잡음 변동시 잡음 억제 능력이 미약하고, 목적 신호의 단구간 진폭 스펙트럼 추정 성능이 낮아진다는 단점을 갖고 있다. 그 이유는 실제 잡음의 스펙트럼은 평균값 주위에 분산되어 있기 때문이 다. 그러므로, 2 채널 Beamformer의 사각(Blocking Matrix)를 이용하여 분석 구간에서의 잡음의 단구간 진폭 스펙트럼을 추정하고, 이 추정된 값을 이용하여 목적 신호의 스펙트럼을 추정하는 기법을 제안하고, 컴퓨터 시뮬레이션을 통하여 그 유효성을 입증한다.

• 한국음향학회지
• /
• 제19권6호
• /
• pp.53-61
• /
• 2000

잡음환경에서의 음성인식 성능향상을 위해서는 서로 다른 잡음환경으로 인한 mismatch를 줄이는 것이 중요하다. 이를 위해 계산이 간단하고 잡음환경에서 비교적 우수한 성능을 내고 있는 스펙트럼 차감법이 널리 사용되고 있다. 본 논문에서는 스펙트럼 차감법을 적용하기 위한 잡음 스펙트럼 추정방법으로 히스토그램 처리방법을 도입한다. 히스토그램 처리방법은 음성이 아닌 구간의 검출이 필요없으며 시간에 따라 변화하는 시변잡음에도 적용 가능한 장점이 있다. 그러나 히스토그램 처리방법으로 신뢰도 높은 잡음 스펙트럼의 평균값을 추정하더라도 스펙트럼 차감법을 적용했을 때의 잔여 잡음의 문제가 발생한다. 이를 해결하기 위하여 잡음추정 과정에 사용되었던 히스토그램의 분포특성을 고려한 새로운 over-estimation 적용방식을 제안한다. 제안된 방식은 측정된 잡음의 분포에 따라 적응적으로 over-estimation의 정도를 결정함으로써 SNR 변화에 따른 영향이 적은 장점이 있다. 자동차 소음 환경에서의 화자독립 고립단어 인식실험 결과, 기존의 over-estimation factor를 적용한 경우보다 제안된 방식의 인식성능이 개선되었다.

• 한국정보전자통신기술학회논문지
• /
• 제6권2호
• /
• pp.81-87
• /
• 2013

Motion segmentation is a fundamental technique for analysing image sequences of real scenes. A process of identifying moving objects from data is a typical task in many computer vision applications. In this paper, we propose motion segmentation that generally consists from background subtraction and foreground pixel segmentation. The Approximated Median Filter (AMF) was chosen to perform background modeling. Motion segmentation in this paper covers RGB video data.

• 한국정보전자통신기술학회논문지
• /
• 제7권1호
• /
• pp.61-67
• /
• 2014

Motion segmentation is a fundamental technique for analysing image sequences of real scenes. A process of identifying moving objects from data is a typical task in many computer vision applications. In this paper, we propose motion segmentation that generally consists from background subtraction and foreground pixel segmentation. The Approximated Median Filter(AMF) was chosen to perform background modeling. Motion segmentation in this paper covers RGB video data.

• 대한의용생체공학회:의공학회지
• /
• 제2권2호
• /
• pp.109-116
• /
• 1981

The objective of this research is to develop the hardware and software systenls of the digital intravenous angiographic system. The system will be used in clinical diagnosis of cardiovascular and cerebrovascular diseases. By utilizing computer image processing technique and digital subtraction method, the system can decrease the patient's harzard and expenses in performing intravenous angiography, as compared with conventional angiograms of the arterial injection.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1987년도 한국자동제어학술회의논문집; 한국과학기술대학, 충남; 16-17 Oct. 1987
• /
• pp.684-689
• /
• 1987

Digital Subtraction Angiography(DSA) technique has been widely used to detect vascular diseases and hemodynamic parameters noninvasively. However, there are few quantitative studies on the factors in fluencing the resultant DSA image quality. In this paper, several important factors are suggested to improved the DSA image quality based on mathematical analysis. Experimental DSA images for different filters are shown and also discussed the difference between original and processed image quality.

• 치과방사선
• /
• 제28권1호
• /
• pp.215-224
• /
• 1998

The following conclusions were obtained from the non-reconstructed and reconstructed subtraction images of the standard intraoral radiographs which were taken with paralleling technique with Rinn XCP only and with occlusal bite registration for geometric standardization using bilateral mandibular premolar and molar regions of two dry human skulls. 1. The SD of the overall subtraction images of the premolars and molars of the non-reconstructed, that is, the manual superimposition showed statistically significant difference between the non-registered and registered groups. 2. In non-reconstructed and non-registered cases, the quality of the subtraction images were improved when superimposition was focally done and this was more evident in areas where the radiographic images tend to be distorted due to anatomic reasons. 3. In non-reconstructed and registered cases, the subtraction images were consistent regardless of the anatomic site or the focus of superimposition. This means that the geometric standardization with only occlusal bite registration could produce serial radiographs which is suitable for subtraction. 4. The SD of the overall subtraction images of the premolars and molars of the reconstructed, that is, the automatic superimposition showed statistically insignificant difference between the non-registered and registered groups. This means that using reconstruction, subtraction radiography is possible without occlusal bite registration. 5. In reconstructed and non-registered cases, compatible quality of the subtraction images were obtained regardless of the anatomic site or area of the corresponding points. 6. In reconstructed and registered cases, best subtraction images whose quality showed sensitivity to the areas of corresponding points were obtained.