http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
RGB 칼라 특징정보와 Median 필터링을 이용한 영상검색
강광원(Gwang-Won Kang),박종안(Jong-An Park) 한국정보기술학회 2007 한국정보기술학회논문지 Vol.5 No.2
This paper presents an image retrieval technique that use RGB color feature information and median filtering to design a more effective algorithm that extracting features of color images for content-based image retrieval. First, extract middle value that use median filtering after segmentation color image into block of regular size. Extraction middle value presents an image retrieval technique using feature after color information do index using difference value and quantization by size order of R, G, B. Compose and use in small feature table by color image feature in this paper that see because need a large amount of storage space and many calculations according to used number of feature at algorithm of effective extract feature. Goodness of fit and order was obtained by comparing of features that were organized into a table, using the proposed algorithm, for the input image and database images. Confirmed that is superior comparing of features that were into a feature table by algorithm that propose and compare input image and DB image and yield fitness and order.
객체 코너의 재배열된 특성요소의 최대 열에 의한 영상 정보 검색
강광원(Gwang-won Kang),이지민(Ji-min Lee),박종안(Jong-an Park) 한국정보기술학회 2010 한국정보기술학회논문지 Vol.8 No.12
본 연구에서는 객체 코너의 재배열된 특성요소의 최대 열에 기반하는 비주얼 정보검색 기법을 제안한다. 제안된 알고리즘은 다음 단계로 진행된다. 먼저 영상 내 객체의 코너 점을 추출하여, 이를 크기순에 따라 순환적으로 재배열하여 특성벡터를 형성한다. 그리고 재배열된 특성벡터에 따라 얻게 되는 최대 열벡터를 기반으로 유사도를 측정하게 된다. 제안된 기법은 영상 내에서 형태 구조가 분명한 객체의 실험에서 성능이 우수하였으고, 객체의 이동이나 회전에도 강인하였으며 코너 패치 히스토그램을 이용한 형태 검색에 비해 약 0.05%의 향상된 recall을 나타내었다. This paper have proposed an algorithm for retrieving a visual information based on rearranged feature vectors of the object corner. The proposed algorithm is processed as follows. First, the corner points of the object in an image are extracted and then the feature vectors are obtained. It are rearranged according to the number dimension and consist of sequence vectors. And the similarity based on the maximum of sequence vectors is measured. The proposed technique is invariant to the rotation or the transfer of the objects and more efficient in case that the objects present simple structure. In simulation that use Wang's database, the method presents that the recall property is improved by 0.05% and more than the standard corner patch histogram.