Abstract:
PURPOSE: A semi-automatic medical image division method and an apparatus thereof are provided to express an area to be modified in real time by applying a high speed processing algorithm based on real time task of a user, thereby improving the accuracy of image division area modification. CONSTITUTION: An area selection calculator(510) calculates an outline part for modification of divided areas of an image through specific image information, initial area information, pointer location information, and, outline range selection information. An area controller(520) applies a preset weight to a proximate area location and modification location section information. The area controller produces the outline part. Based on selection operation of the outline part, the area controller decides the range of exact areas through the control of the divided areas.
Abstract:
Disclosed are an apparatus and a method for region growing which perform operations related to a region growing technique in parallel using multiple cores which are connected over a network. The apparatus for region growing using multiple cores comprises: an operation unit for performing operations regarding region growing of a two-dimensional or three-dimensional pixel region; multiple cores including an inner memory unit for storing a queue related to a seed pixel which is used for the operations; and a shared memory unit which is connected to the multiple cores over a network and shared by the multiple cores. The method for region growing using multiple cores comprises the following steps: recording a core flag which indicates the operating state of each core in a core flag recording area in the shared memory unit; determining whether an operation for region growing has been performed with regard to at least one pixel adjacent to a seed pixel which is inputted to a queue in a core whose operating state is active among the multiple cores based on a test bit value of the pixel flag corresponding to the adjacent pixel; changing the test bit value of the adjacent pixel if the operation for region growing has not performed with regard to the adjacent pixel; performing the operation for region growing which determines whether the adjacent pixel is included in a region of the seed pixel; and recording the operation result for region growing in an inclusion bit value of the pixel flag corresponding to the adjacent pixel.
Abstract:
PURPOSE: A volume maintenance area detection and non rigid registration method using GLCM(Grey Level Co-occurrence Matrix) texture analysis and a recording medium thereof are provided to accurately detect a volume maintenance area. CONSTITUTION: An optimum rigid transform image with optimum similarity is generated(S100). A volume maintenance area is classified(S230). A transformation parameter for B-spline transformation is calculated(S310). A floating image is converted using the B-spline transformation(S320). The transform parameter for the B-spline conversion is updated (S340). A registered floating image with the minimum cost function is generated(S350).
Abstract:
본 발명은 별도의 악기나 미디 장비를 통하지 않고서도 사용자가 입력하는 음성을 분석하여 작곡이 이루어진 악보를 생성할 수 있고, 그 음성 분석 정보를 근거로 기성곡을 검색하여 원하는 곡을 찾을 수 있도록 하는 입력 음성의 분석을 이용한 작곡 및 기성곡 검색 시스템을 제공한다. 이를 위해 본 발명은 사용자로부터 입력되는 음성을 잡음 제거 및 디지털 변환을 거쳐서 저장하는 작곡 음성 입력 장치와, 해당 시스템의 실행 및 구동 상태를 설정하는 키입력을 수행하고, 사용자의 음성 분석에 의해 생성된 악보의 내용을 수정, 삭제, 추가하기 위한 키입력을 수행하는 키입력부, 사용자의 입력 음성 데이터를 해밍 윈도우를 이용하여 FFT 및 자기 상관 함수를 이용한 함수 연산 처리 알고리즘을 통해 분석하고, 그 분석된 음에 대응하는 악보 기호를 생성하여 악보 정보를 생성함으로써, 상기 키입력부의 키입력에 따라 악보 정보의 디스플레이와, 악보 편집, 프린터 출력이 가능하도록 하는 작곡 처리 장치, 상기 완성된 악보 정보에 포함된 음정과 박자 정보를 근거로, 다수의 기성곡 중에서 가장 유사한 음정 및 박자 정보를 갖는 기성곡을 검색하여 추출하는 기성곡 검색 처리 장치 및, 다수의 기성곡 정보가 악보 스트링 정보 형태로 저장되어 상기 기성곡 검색 장치로 기성곡의 검색을 위해 제공하는 데이터베이스를 포함하여 구성된 것을 특징으로 한다.
Abstract:
PURPOSE: An apparatus and a method for filtering geographical data are provided to filter search target with a threshold value using a ladder type interface. CONSTITUTION: An interface processor(120) receives the selection of a map about the search target area. The interface processor receives the search target and search condition, and outputs the ladder type interface. A filtering unit(112) confirms a threshold value corresponding to the height value of the ladder type interface by the part of the map. The filtering unit filters the search target satisfying the threshold value, and extracts the filtered value. A screen configuration part(114) displays the filtered value to the map.