• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10a
• /
• pp.127-129
• /
• 2000

일반적인 색 정보추출방법으로 색 히스토그램(Color Histogram)은 색의 분포나 응집성, 질감에 대한 구분능력이 없다는 단점을 가지고 있어 정환한 이미지 유사성 비교를 위해 추가적인 정보를 요구한다. Androutsos등은 Haar Wavelet 변환을 통해 이미지의 방향성 질감정보를 구하였다[1]. 하지만 이 방법은 Haar Wavelet 변환의 특성으로 인해 정확한 방향성 정보를 얻을 수 없었다. 본 논문에서는 인접 픽셀(pixel)값의 편차(deviaiton)를 이용하여 방향성 정보를 추출 성능을 향상시키는 방법을 제안하였고, Brodatz 112 질감 이미지와 실재 자연사진을 통해 방향성 질감의 성능을 평가하였다.

• FLOWER RESEARCH JOURNAL
• /
• v.19 no.4
• /
• pp.238-244
• /
• 2011

The study was conducted in order to systematize specialization of flower decoration. The results of this study are summarized as follows. The research of image adjectives were analyzed into a total of 4 factors, which are texture, color, period, decoration. With regard to texture, the flower with the highest texture of 2.00 was white Zantedeschia aethiopica and that with the lowest texture of -1.90 was Leucospermum cordifolium. In regard of color, the flower with the highest color of 1.70 was Rosa 'Judy' and that with the lowest color of -2.40 was Gypsophila. With regard to period, the flower with the highest period of 2.00 was Gomphrena globosa and that of the lowest of -3.00 was Leucospermum cordifolium. In decoration, the flower with the highest decoration of 2.50 was Agapanthus 'Intermedia' and that with the lowest decoration of -2.00 was Heliconia bihai 'Lobster'. As for correlation among the factors, color and period were in a correlation with each other. It shows that color and period are the most important factors in flower decoration. In result of conjoint analysis, importance was high in order of color, texture, decoration and period. This suggests that consumers overemphasize color and texture in flower decoration and may fail to notice the importance of period.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.06a
• /
• pp.169-172
• /
• 2001

영상의 특성 파악을 위해서 질감 특징이 많이 사용되고 있다 Co-occurrence matrix를 이용한 질감은 영상의 변화형태에 대한 수치자료로 다양한 함수들을 가지고 있으며, 영상의 특성에 따라서 그 함수들을 활용하여 영상의 분할과 분류에 사용하고 있다. 본 논문에서는 질감 특징을 시각화하기 위한 방법으로 GLCM의 로컬값을 새로운 픽셀값으로 하는 영상화 기법에 대해 논하였다. 실험을 통해 질감특징 중 대조적인 관계와 동일성을 가진 질감에 대한 영상을 얻을 수 있었으며, 영상 분석에 대한 시각적인 자료를 얻을 수 있었다. 질감특징은 각 항수별 특징값의 효율적인 사용을 위해 시각화되어질 필요성이 있으며 영상화되어진 질감특징영상을 이용하면 영상의 분석과 이해에 효과적인 접근이 가능하다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.11a
• /
• pp.447-452
• /
• 2001

내용 기반 이미지 검색에서 질감정보는 이미지의 검색 속성으로 사용할 수 있는 중요한 정보를 가지고 있다. 본 논문에서는 검색의 이미지 속성으로서 질감 특징을 사용한다. 의료영상 MRI 중 특히 뇌영상의 검색에서 질감의 특징은 전체 이미지를 대상으로 한 전역 질감 특징 값과 종양이나 뇌출혈 부분 등 정상이 아닌 이상객체 부분의 지역 질감 특징 값을 3단계 웨이블릿 변환을 통해 추출하고 추출된 여러 개의 특징 중 검색 효율성을 높일 수 있는 특징만을 선별하여 검색에 이용하는 방안을 제안한다.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.4
• /
• pp.34-43
• /
• 2010

Texture information plays an important role in object recognition and classification. To perform an accurate classification, the texture feature used in the classification must be highly discriminative. This paper presents a novel texture descriptor for texture-based image retrieval and its application in Computer-Aided Diagnosis (CAD) system for Emphysema classification. The texture descriptor is based on the combination of local surrounding neighborhood difference and centralized neighborhood difference and is named as Combined Neighborhood Difference (CND). The local differences of surrounding neighborhood difference and centralized neighborhood difference between pixels are compared and converted into binary codewords. Then binomial factor is assigned to the codewords in order to convert them into high discriminative unique values. The distribution of these unique values is computed and used as the texture feature vectors. The texture classification accuracies using Outex and Brodatz dataset show that CND achieves an average of 92.5%, whereas LBP, LND and Gabor filter achieve 89.3%, 90.7% and 83.6%, respectively. The implementations of CND in the computer-aided diagnosis of Emphysema is also presented in this paper.

• Journal of KIISE:Software and Applications
• /
• v.34 no.10
• /
• pp.900-909
• /
• 2007

This paper describes an experiment on the evaluation of human sensibility by monitoring responses to changes In the frequency and amplitude of a tactile display system. Preliminary tasks were performed to obtain effective adjectives concerning texture perception. The number of collected adjectives was originally 33. This number of adjectives was reduced to 14 by a suitability survey that asked whether an adjective is suitable for expressing a texture feeling. Finally after performing a semantic similarity evaluation, the number of adjectives was further reduced to ten and these ten were used in the main experiment. In the main experiment, selected sandpaper types and 15 selected combinations of frequencies and amplitudes of a tactile display were utilized to quantitatively evaluate the ten adjectives using a bipolar seven-point scale. The data show that a relationship exists between the independent variables(frequency, amplitude, and grit site) and the dependent variable(perceived texture). That is, the change of frequency and amplitude is directly related to perceived roughness or essential elements of human tactile sensitivity found in the preliminary experiment.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.10
• /
• pp.83-92
• /
• 2011

The aim of this study is to evaluate whether 3D nuclear chromatin texture features are significant in recognizing the progression of cervical cancer. In particular, we assessed that our method could detect subtle differences in the chromatin pattern of seemingly normal cells on specimens with malignancy. We extracted nuclear texture features based on 3D GLCM(Gray Level Co occurrence Matrix) and 3D Wavelet transform from 100 cell volume data for each group (Normal, LSIL and HSIL). To evaluate the feasibility of 3D chromatin texture analysis, we compared the correct classification rate for each of the classifiers using them. In addition to this, we compared the correct classification rates for the classifiers using the proposed 3D nuclear texture features and the 2D nuclear texture features which were extracted in the same way. The results showed that the classifier using the 3D nuclear texture features provided better results. This means our method could improve the accuracy and reproducibility of quantification of cervical cell.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10b
• /
• pp.362-364
• /
• 2000

본 본문에서는 간단하면서 효과적인 관련성 귀환을 가진 영상 검색시스템에 대하여 연구하였다. 먼저 영상 데이터베이스 내에 있는 영상들에 대하여 Gabor Wavelet 변환을 이용하여 질감특징을 추출하고, 추출한 특징값을 다양한 형태로 영상검색에 이용하였다. 초기 검색결과에 대하여 관련성 귀환을 영상 검색시스템에 적용하고, 이를 기존의 관련성 귀환을 가진 시스템과 비교하였다. 16종류의 512개의 영상으로 구성된 영상 데이터베이스에 대하여 실험한 결과, 제한된 방법은 INRIA의 방법보다 각 귀환단계에서 약 7~8%의 높은 검색 효율을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2001.04b
• /
• pp.863-866
• /
• 2001

본 논문에서는 관련성 귀환을 가진 칼라와 질감기반의 영상검색 시스템에 대하여 연구하였다. 먼저 영상데이터베이스 내에 있는 영상들에 대하여 칼라특징, 질감특징을 추출하고 추출된 특징 값을 다양한 형태로 영상검색에 이용하였다. 그리고 초기 검색결과에 대하여 사용자 평가를 관련성 귀환을 통하여 영상검색 시스템에 적용하고, 개선된 결과를 얻었다. 16종류의 다양한 영상으로 구성된 영상 데이터베이스에 대하여 실험한 결과, 제안된 방법은 INRIA의 방법보다 각 귀환단계에서 약 10%$\sim$l6% 이상의 높은 검색율을 보였다.

• Journal of Korea Society of Industrial Information Systems
• /
• v.13 no.4
• /
• pp.108-114
• /
• 2008

In this paper, we proposed robust blind watermarking algorithm using texture block coefficient based on discrete cosine transform(DCT). The proposed method embedded all of watermark signals into DC component of $8\times8$ block DCT in order to robust various external attack The texture block coefficient was composed absolute value of DCT coefficients. Experiment showed that the proposed method better than conventional methods in the invisibility and various attack such as dithering, cropping, and scaling.