• 제목/요약/키워드: wavelet coefficient differences

검색결과 4건 처리시간 0.019초

이웃한 웨이브릿 변환 계수 쌍의 평균과 차이를 이용한 워터마킹 기법 (A Watermarking Technique Using Means and Differences of Neighboring Wavelet Transform Coefficient Pairs)

  • 김현순;배성호;박길흠
    • 한국정보처리학회논문지
    • /
    • 제7권6호
    • /
    • pp.1980-1987
    • /
    • 2000
  • In this paper, an efficient watermarking technique in wavelet transform domain is proposed. Watermarking is embedding a digital signal called as 'watermark' into images to claim the ownership. In the proposed method, the image is 1-level wavelet transformed, and then the watermark with a binary stamp is embedded into the baseband. The watermark is embedded by inverting the polarities of he selected coefficient paris. In the inverting process, we can increase image quality by finding means and differences of the selected neighboring coefficient paris, and then adding values, which are inversely proportional to the differences, to th means. The experimental results show that the proposed method has good quality and is robust to various attacks such as the JPEG lossy comparison, noise addition, clipping, blurring, etc.

  • PDF

Damage detection for a beam under transient excitation via three different algorithms

  • Zhao, Ying;Noori, Mohammad;Altabey, Wael A.
    • Structural Engineering and Mechanics
    • /
    • 제64권6권
    • /
    • pp.803-817
    • /
    • 2017
  • Structural health monitoring has increasingly been a focus within the civil engineering research community over the last few decades. With increasing application of sensor networks in large structures and infrastructure systems, effective use and development of robust algorithms to analyze large volumes of data and to extract the desired features has become a challenging problem. In this paper, we grasp some precautions and key points of the signal processing approach, wavelet, establish a relative reliable framework, and analyze three problems that require attention when applying wavelet based damage detection approach. The cases studies how to use optimal scales for extracting mode shapes and modal curvatures in a reinforced concrete beam and how to effectively identify damages using maximum curves of wavelet coefficient differences. Moreover, how to make a recognition based on the wavelet multi-resolution analysis, wavelet packet energy, and fuzzy sets is a meaningful topic that has been addressed in this work. The relative systematic work that compasses algorithms, structures and evaluation paves a way to a framework regarding effective structural health monitoring, orientation, decision and action.

Watermarking for Digital Images Using Differences and Means of the Neighboring Wavelet Coefficients

  • Kim, Hyun-Soon;Bae, Sung-Ho;Yoon, Ock-Kyung;Park, Kil-Houm
    • 대한전자공학회:학술대회논문집
    • /
    • 대한전자공학회 2000년도 ITC-CSCC -1
    • /
    • pp.466-469
    • /
    • 2000
  • In this paper, a watermarking technique for digital images is proposed. In our method, an image is 1-1eve1 wavelet transformed, and then the watermark of a binary stamp is embedded into the baseband. The watermark is embedded by inverting the polarities of the selected coefficient pairs. In the inverting process, we can increase perceptual image quality by finding means and differences of the selected neighboring coefficient pairs, and then adding values, which are inversely proportional to the differences, to the means. The experimental results show that the proposed method has good quality and is robust to JPEG lossy compression and various image processing operations.

  • PDF

산지토양의 탄소와 질소 예측을 위한 가시 근적외선 분광반사특성 분석의 전처리 방법 비교 (Evaluating Spectral Preprocessing Methods for Visible and Near Infrared Reflectance Spectroscopy to Predict Soil Carbon and Nitrogen in Mountainous Areas)

  • 정관용
    • 대한지리학회지
    • /
    • 제51권4호
    • /
    • pp.509-523
    • /
    • 2016
  • 토양 예측은 지속가능한 산지관리 측면에서 필요한 토양특성자료를 제공할 수 있다. 이중 가시 근적외선 분광반사 특성을 이용한 토양 예측은 저비용, 빠른 분석과 비파괴 측정, 비교적 높은 정확도로 관심을 받고 있다. 일반적으로 토양 분광반사특성 측정 과정에서 잡음이 나타날 수 있어 전처리 과정이 필요하다. 하지만 이러한 전처리 방법을 비교하고 평가하는 작업이 거의 이루어지지 못 했다. 본 연구에서는 토양 탄소와 질소 예측을 위해 5가지 전처리 방법을 비교하였다. 이는 연속체 제거, Savitzky-Golay 변환, 이산 웨이블렛(wavelet) 변환, 1차와 2차 도함수 변환이다. 토양예측 모델로 부분 최소제곱 회귀모형을 사용하였고, 총 153개 시료 중에서 검증을 위해 122개 훈련자료와 31개의 검증자료로 나누어 평가하였다. 전반적으로 토양시료의 탄소 함량이 높을수록 토양에 대한 입사 에너지의 흡수가 커지는 특성을 보였다. 파장별로는 가시광선 영역(650nm와 700nm)이 토양 탄소 그리고 질소와 가장 높은 상관관계를 보였다. 전처리 비교에서 연속체 제거가 토양 탄소(9.53mg/g)와 질소(0.79mg/g)에 대해 가장 높은 정확도(Root Mean Square Error)를 보였다. 따라서 토양 탄소와 질소 예측을 위해 연속체 제거가 가장 효과적인 분광반사특성 분석의 전처리 방법으로 판단되었다. 시각적인 평가에서 웨이블릿 변환이나 Savitzky-Golay 변환은 차이가 거의 없었고, 평가 결과도 유사했다. 따라서 다소 계산과정이 간단한 Savitzky-Golay 변환이 선호될 수 있다.

  • PDF