Abstract:
본 발명은 입체영상촬영장치에서의 주시각 제어장치에 관한 것으로서, 스테레오 카메라에 의하여 스테레오 영상을 구성하는 시차(視差)가 다른 2개의 이미지를 획득하는 촬영부(100); 상기 획득된 2개의 이미지의 시차로부터 상기 스테레오 영상의 변이 정보인 디스패리티(Disparity) 정보를 추출하는 영상 처리부(300); 상기 추출된 디스패리티 정보를 이용하여 디스패리티 맵(Disparity Map)을 생성하고, 상기 디스패리티 맵을 이용하여 상기 스테레오 영상에 포함된 객체들 각각에 대한 이미지의 깊이를 나타내는 제1 뎁스맵(Depth-Map)을 생성하는 깊이 영역 획득부(400); 상기 생성된 제1 뎁스맵을 히스토그램(Histogram)화 하고 상기 히스토그램을 이용하여 상기 객체들 각각을 거리별로 구분하는 제2 뎁스맵을 생성하는 세그먼테이션부(450); 상기 생성된 제2 뎁스맵으로부터 상기 객체들 각각에 대한 VFL(Visual Fatigue Level)값을 산출하는 VFL 획득부(500); 상기 산출된 VFL값 중 가장 높은 값을 갖는 객체를 초점으로 주시각을 자동 제어하고, 주시각이 자동 제어된 스테레오 영상을 디스플레이에 출력하는 자동 주시각 제어부(600)를 포함한다. 이에 의하면, 스테레오 카메라 촬영 후 짧은 시간 내의 간단한 조작만으로도 VFL값에 따라 사용자가 원하는 객체에 초점을 맞추어 주시각을 제어할 수 있는 효과가 있다.
Abstract:
PURPOSE: A convergence angle control device of a stereoscopic imaging device is provided to control a convergence angle by focusing on a desired object according to a visual fatigue level (VFL) value with only a simple operation within short time after taking an image with a stereoscopic camera. CONSTITUTION: A depth region obtaining part (400) generates a disparity map by using disparity information extracted from an image processing part (300) and generates a first depth map by using the disparity map. A segmentation part (450) converts the generated first depth map into a histogram and generates a second depth map by using the histogram. A VFL obtaining part (500) calculates a VFL value for each object from the second depth map. An automatic convergence angle control part (600) automatically controls a convergence angle by focusing on an object having the highest VFL value among the calculated VFL values. [Reference numerals] (101) Left camera; (102) Right camera; (200) Image obtaining part; (300) Image processing part; (400) Depth region obtaining part; (450) Segmentation; (500) VFL obtaining part; (600) Automatic convergence angle control part (Apply a VFL value); (700) Semiautomatic convergence angle control part; (800) User input; (850) Display part
Abstract:
PURPOSE: A Korean alphabet input method in a portable terminal is provided to eliminate additional space key input process by changing a consonant inputted with an initial sound into the end sound of a previous syllable. CONSTITUTION: When characters are inputted in a Korean alphabet input mode, a current input character is determined a consonant or a vowel(S1). In case the current input character is determined as the consonant, the initial sound of the current input character is extracted according to kinds of characters which are just before being inputted(S2). In case the current input character is determined as the vowel, the current input character is outputted(S3).
Abstract:
A stereo matching apparatus which generates a displacement map based on stereo images according to embodiments of the present invention comprises: a displacement map calculation unit for generating a full-scale displacement map regarding the stereo images in a predicted search area for a current frame generated from a previous frame; an effective search area calculation unit which sets an object area extracted by comparing a histogram log scale of the full-scale displacement map with a threshold value as an effective search area of the current frame; and an predicted search area calculation unit which generates the predicted search area for the next frame based on at least one linear transformation element extracted among the histograms of the effective search areas of the previous frame and the current frame. [Reference numerals] (AA) Start; (BB) k=1 or M; (S10) Each receive stereo images of left and right components from a current frame; (S11) k = 1,2 or M?; (S20) Generate a full-scale displacement map and an effective search area by a simple search area method(SSR); (S30) Generate a full-scale displacement map, an effective search area, and a predicted search area in the effective search area; (S40) Generate a full-scale displacement map, an effective search area, and a predicted search area in the predicted search area