| 摘要 |
A method of compressing a frame of pixel data comprises receiving a frame of pixel data S1; performing a transform of the pixel data comprising: replacing adjacent pixels (A, B) by the sum (A+B) and difference (A-B) S2 of the pixel data; in a binary representation of the sum (A+B) and difference (A-B), discarding S3 the lowest bit of one of the sum (A+B) and difference (A-B); repeating S4 the transform for the entire frame of pixel data; and then performing entropy coding of transformed pixel data, S5. The frame of pixel data may be a frame of pixel tiles, with transform and entropy coding being performed on a tile by tile basis (Figure 1, S2, with colour transforming, S1). The step of performing a transform for the entire pixel data frame may be performed twice for a frame, once (first) using horizontally adjacent pixels and once (second) using pixels adjacent vertically. Thus, the method adapts the known Haar wavelet transform - which also uses replacement of adjacent pixels with their sums and differences â by discarding the lowest bit of the sum or difference, providing a greater level of compression and computational simplicity. |
