| 主权项 |
1. A method implemented in a multimedia conversion system for converting two-dimensional (2D) multimedia data to stereoscopic multimedia data, comprising:
determining vertical characteristics of pixels in a frame of the 2D multimedia data, wherein determining vertical characteristics of the pixels comprises for each scan line in the frame, determining a difference between pixel pairs comprising a pixel in a current scan line and a pixel in a second scan line, wherein the vertical characteristics comprise at least one of: a scan line absolute difference; a difference between pixels in the second scan line; and an absolute difference between pixels in the second scan line, wherein determining vertical characteristics further comprises determining a running sum for each position y, wherein determining a running sum is calculated according to the following:Sum[y]=∑i=1ymax(w1*D1[i]+w2*D2[i]+w3*D3[i],lower_bound), wherein Sum re resents the running SUM D1[i] represents the scan line absolute difference, D2[i] represents the difference between pixels in the second scan line, D3[i] the absolute difference between neighboring pixels in the second scan line, max ( ) represents a maximum function, lower bound is a constant ranging from 0 to negative infinity, and w1, w2, w3 are weighting factors of D1, D2, D3; calculating a depth for each pixel in the 2D multimedia data according to a depth of a vertical pixel in the second scan line and the determined vertical characteristics of the pixel; and generating a complementary stereoscopic image pair according to the depth of each pixel. |
