Line-based video compression
A method for line-based video rate control is provided. The line based video rate control method includes system feedback to change system operating parameters, including on a packet-by-packet basis and also on a line-by-line basis. Also provided is a method for line-based compression. The method includes basic elements of an arithmetic coder to improve Golomb coding performance. By inverting operations in the method for line-based compression, the corresponding decoder can be obtained. The method also provides a heuristic-driven method for generating prediction residuals from quantized data, wherein prediction is driven in the direction of maximum correlation so that it is unnecessary to supply the decoder with additional data specifying this direction at each array element.
1. A method for line based video compression, comprising:
selecting a video format based on line of a video frame communicated between two nodes of said video frame;
selecting a wavelet transform based on said line of said video frame;
providing a low-pass portion and a high-pass portion of said line of said video frame based on said selected wavelet transform;
applying quantization parameters to said low-pass portion and to said high-pass portion of said line of said video frame to generate a quantized low-pass portion and a quantized high-pass portion; and
conditioning said quantized low-pass portion to fit an exponential distribution based on a low-pass portion of a previous line of said video frame.
2. The method of claim 1 , further comprising encoding said conditioned low-pass portion and said quantized high-pass portion.
3. The method of claim 1 wherein said video format includes one of an RGB and a YUV format.
4. The method of claim 3 wherein said YUV format includes one of a 4:4:4, 4:2:2, and 4:1:1 format.
5. The method of claim 1 wherein selecting said wavelet transform includes selecting one of a Daubechies-(7,9), and LeGall-(5,3) wavelet transforms.
6. The method of claim 1 wherein selecting said wavelet transform includes selecting said wavelet transform based on an image compression standard.
7. The method of claim 2 wherein said encoding includes using an encoder that selectively switches between arithmetic coding and Golomb coding based on said active line of said video frame.
8. The method of claim 7 wherein encoding includes using a Golomb encoder implemented without arithmetic.
9. The method of claim 1 further comprising:
predicting a vertical edge in said high-pass portion;
obtaining a residual high-pass portion by suppressing said vertical edge in said high-pass portion; and
encoding said residual high-pass portion.
10. The method of claim 9 wherein predicting said vertical edge in said high-pass portion includes processing said low-pass portion with a vertical edge detector and wherein suppressing said vertical edge in said high-pass portion includes subtracting said predicted vertical edge from said high-pass portion.
11. The method of claim 1 further comprising:
generating at least one high-pass wavelet coefficient; and
generating at least one low-pass wavelet residual.
12. The method of claim 11 further comprising:
encoding at least one high-pass wavelet coefficient; and
encoding at least one low-pass wavelet residual.
13. The method of claim 12 wherein encoding said at least one high-pass wavelet coefficient includes Golomb encoding.
14. The method of claim 12 wherein encoding said at least one low-pass wavelet residual includes Golomb encoding.
15. A line based video compression apparatus, comprising:
means for selecting a video format based on a line of a video frame;
means for selecting a wavelet transform based on said line of said video frame;
means for providing a low-pass portion and a high-pass portion of said line of said video frame based on said selected wavelet transform;
means for applying quantization parameters to said low-pass portion and to said high-pass portion of said line of said video frame to generate a quantized low-pass portion and a quantized high-pass portion; and
means for conditioning said quantized low-pass portion to fit an exponential distribution based on a low-pass portion of a previous line of said video frame.
16. The apparatus of claim 15 , further comprising means for encoding said conditioned low-pass portion and said quantized high-pass portion.
17. The apparatus of claim 16 , wherein said means for encoding includes an encoder that selectively switches between arithmetic coding and Golomb coding based on said line of said video frame.
18. The apparatus of claim 15 , wherein said video format includes one of an RGB and a YUV format.
19. The apparatus of claim 15 , further comprising:
means for predicting a vertical edge in said high-pass portion;
means for obtaining a residual high-pass portion by suppressing said vertical edge in said high-pass portion; and
means for encoding said residual high-pass portion.
20. The apparatus of claim 15 , further comprising:
means for generating at least one high-pass wavelet coefficient; and
means for generating at least one low-pass wavelet residual.
21. The apparatus of claim 20 , further comprising:
means for encoding at least one high-pass wavelet coefficient; and
means for encoding at least one low-pass wavelet residual.