Method and apparatus for progressive encoding for text transmission
A method for encoding a text image. In one embodiment the method comprises determining a background color of an image; dividing the image into text pixels and background pixels, each of the background pixels comprising the background color; determining a text color based on colors of the text pixels; determining a text attribute comprising a luminance profile describing a contour of luminance of a set of the text pixels; generating a binary text mask based on position and luminance of each of the text pixels and the luminance profile; and generating, by a computer, an encoded image comprising the background color, the text color, the binary text mask, and the text attribute.
1. A method for encoding a text image, comprising:
determining, by a computer having one or more processors, a background color of an image;
dividing, by the computer, the image into text pixels and background pixels, each of the background pixels comprising the background color;
determining, by the computer, a text color based on colors of the text pixels;
determining, by the computer, a text attribute comprising a luminance profile describing a contour of luminance of a set of the text pixels;
generating, by the computer, a binary text mask based on position and luminance of each of the text pixels and the luminance profile; and
generating, by the computer, an encoded image comprising the background color, the text color, the binary text mask, and the text attribute.
2. The method of claim 1 , further comprising decomposing a frame buffer image, rendered by the computer, into an unchanged portion, a text portion and a picture portion and selecting the image from the text portion, wherein decomposing the frame buffer image comprises filtering the frame buffer image to identify the text portion.
3. The method of claim 1 , wherein the background color comprises a single set of a red color value, a green color value and a blue color value and determining the background color comprises determining pixel density information of the background color in the image.
4. The method of claim 1 , wherein dividing the image comprises analyzing color planes of the image and dividing the image into the text pixels and the background pixels on sub-pixel boundaries between red, green and blue pixel elements.
5. The method of claim 1 , wherein the text color comprises two chrominance values and a luminance value, and wherein the text color is representative of all the text pixels.
6. The method of claim 1 , wherein determining the text attribute comprises comparing profiles of pixel luminance of a plurality of segments of the text pixels to a plurality of reference pixel luminance profiles, the plurality of reference pixel luminance profiles representative of at least one of block text, bold text or anti-aliased text.
7. The method of claim 1 , wherein generating the binary text mask comprises identifying a pattern of pixels that, when filtered by a filter associated with the text attribute, generates an approximation of a luminance pattern of the text pixels.
8. The method of claim 7 , wherein the pattern of pixels comprises a defined quantity limit of pixel luminance levels, wherein the defined quantity limit is adjusted based on a desired compression ratio between the image and the encoded image.
9. The method of claim 7 , wherein generating the binary text mask comprises modifying the binary text mask to improve a prediction of an encoder used in generating the encoded image.
10. The method of claim 1 , further comprising generating at least one refinement mask comprising up to eight levels of luminance level information for each of a selection of the text pixels, and encoding the at least one refinement mask to generate an encoded refinement.
11. The method of claim 10 , further comprising transmitting, via an inter-computer network, the encoded image and the encoded refinement, wherein the encoded refinement is transmitted as (i) part of the encoded image or (ii) as a subsequent transmission to the encoded image.
12. The method of claim 1 , further comprising decoding the encoded image to generate a decoded image, wherein the decoded image comprises (i) decoded background pixels comprising the background color and (ii) decoded text pixels, each of the decoded text pixels comprising a decoded luminance value determined by the binary text mask and the luminance profile.
13. The method of claim 1 , wherein the background color is a first single color value, the text color is a second single color value, the binary text mask comprises a single bit per pixel of the image, and the text attribute is a single value used to define the luminance profile.
14. The method of claim 1 , further comprising generating a refinement encoding comprising at least one of an additional binary text mask plane, an additional text color or additional text attribute information.
15. The method of claim 1 , wherein generating the encoded image comprises encoding an indication of repeating, from a previous image, at least one of the background color, the text color or the text attribute.
16. The method of claim 1 , further comprising converting pixels of the image from a red, green and blue color space to a chrominance and luminance color space.