Technique for determining the slope of a field pixel
View Patent ↗A technique for determining the slope of a field pixel is described. According to the technique, one or more sets of diagonal field pixels are downscaled (or downsampled) before they are provided to respective edge detector circuits. By downscaling the sets of diagonal field pixels before they are provided to respective detector circuits, the edge detector circuits detect diagonal edges and, in particular, shallow diagonal edges with greater accuracy. As such, the slopes assigned to the field pixels are more likely to be correct. This ultimately results in a high quality progressive video signal that can be used to generate an image that is completely or substantially free from objectionable artifacts.
1. A method of obtaining an upscaled version of an image, the method comprising:
downscaling the image;
detecting slope for one or more edges in the downscaled version of the image;
determining interpolated pixels for an upscaled version of the image based on the detected slope; and
generating the upscaled version of the image using the interpolated pixels, wherein downscaling comprises downscaling in a first direction while maintaining the scaling in a second direction.
2. The method of claim 1 , wherein the first direction is the horizontal direction and the second direction is the vertical.
3. The method of claim 1 , wherein the image is a frame of an interlaced video signal and the upscaled version is a progressive video signal.
4. The method of claim 1 , wherein the image is from an MPEG video stream.
5. The method of claim 1 , wherein the image is an interlaced video signal and the upscaled version of the image is for a high-definition television signal.
6. A method of obtaining an upscaled version of an image, the method comprising:
downscaling the image;
detecting slope for one or more edges in the downscaled version of the image;
determining interpolated pixels for an upscaled version of the image based on the detected slope; and
generating the upscaled version of the image using the interpolated pixels, wherein detecting comprises:
high-pass filtering the downscaled version of the image to generate a detected value; and
comparing the detected value to a threshold value.