Abstract: An edge-preserving filtering method and apparatus. The edge-preserving filtering method includes: acquiring an upper edge pixel P(x,y?n), a lower edge pixel P(x,y+m), (m+n) left edge pixels P(x?p,y?), and (m+n) right edge pixels P(x+q,y?) of a pixel to be filtered P(x,y), wherein “x”, “y”, “m”, “n”, “p” and “q” are all positive integers, and “y” takes an integer larger than “y?n” and less than “y+m” in turn; acquiring a filtered pixel value of the pixel to be filtered by using pixel values of the pixel to be filtered, the upper edge pixel, the lower edge pixel, the left edge pixels and the right edge pixels. The edge-preserving filtering method and apparatus in accordance with an embodiment utilize edge information around the pixel to be filtered to swiftly acquire a set of pixels for filtering the pixel to be filtered, and thus implement edge-persevering filtering.

Abstract: A method and system for stereo correspondence. The method for stereo correspondence includes a matching cost computation step, a cost aggregation step, a disparity computation step, and a disparity optimization step. The matching cost computation step acquires a left disparity space image and a right disparity space image by using horizontal gradients and vertical gradients of intensities of all component channels of every pixel in a left image and a right image. Utilizing the invention, accurate disparity maps may be acquired quickly.

Abstract: Image defogging method and system, the method including: constructing a pixel-based dark channel map of a fogging image; constructing a local area-based dark channel map of the image; acquiring a final dark channel map of the image; acquiring intensity values of R, G, B channels of a pixel having a maximum grey value of all pixels in an area, covered by a brightest area in the local map, in the fogging map, as R, G, B components of an air light value of the image; acquiring a transmission map of the image by using the final map, a maximum value of the components of the air light value, and a defogging parameter; and acquiring intensity values of the channels of pixels in a defogged image by using the transmission map, the air light value, and the intensity values of the R, G, B channels of the pixels in the image.

Abstract: A method and system for stereo correspondence. The method for stereo correspondence includes a matching cost computation step, a cost aggregation step, a disparity computation step, and a disparity optimization step. The matching cost computation step acquires a left disparity space image and a right disparity space image by using horizontal gradients and vertical gradients of intensities of all component channels of every pixel in a left image and a right image. Utilizing the invention, accurate disparity maps may be acquired quickly.

Abstract: An edge-preserving filtering method and apparatus. The edge-preserving filtering method includes: acquiring an upper edge pixel P(x,y?n), a lower edge pixel P(x,y+m), (m+n) left edge pixels P(x?p,y?), and (m+n) right edge pixels P(x+q,y?) of a pixel to be filtered P(x,y), wherein “x”, “y”, “m”, “n”, “p” and “q” are all positive integers, and “y” takes an integer larger than “y?n” and less than “y+m” in turn; acquiring a filtered pixel value of the pixel to be filtered by using pixel values of the pixel to be filtered, the upper edge pixel, the lower edge pixel, the left edge pixels and the right edge pixels. The edge-preserving filtering method and apparatus in accordance with an embodiment utilize edge information around the pixel to be filtered to swiftly acquire a set of pixels for filtering the pixel to be filtered, and thus implement edge-persevering filtering.

Abstract: A method and apparatus for traversing a triangle in graphic rasterization are provided. The method includes creating a bounding box which is the smallest rectangle consisting of pixels that covers the triangle; choosing a traversal start pixel for traversing the triangle based on the bounding box; choosing a traversal direction based on the position relation between the traversal start pixel and the triangle; and traversing the triangle from the traversal start pixel in the traversal direction. The traversing the triangle includes calculating a next scan start pixel based on slope of a edge function of the triangle with respect to the traversal start pixel and the traversal direction.