Psychovisual image compression
Psychovisual image compression techniques are disclosed that compress pixel data by a fixed compression ratio with little or no perceptual loss of detail. In some implementations, a psychovisual compression process is selected among several psychovisual compression processes based on characteristics of the pixel data. Compression is achieved during encoding by discarding psychovisually unnecessary bits from the pixel data. The psychovisual compression processes can be implemented in hardware and operate on scan lines of pixels captured by the image sensor. The psychovisual compression techniques can be used with image compression techniques to compress further the pixel data.
1. A method comprising:
obtaining pixel data having a first data size;
determining a processing mode from a number of available processing modes based on characteristics of the pixel data and one or more simplex boundaries and exclusion in a color subspace;
selecting a psychovisual compression process based on the determined processing mode; and
psychovisually compressing the pixel data to a second data size using the selected psychovisual compression process, where the second data size is less than the first data size and the first data size and the second data size have a fixed ratio,
where the method is performed by at least one hardware apparatus.
2. The method of claim 1 , further comprising:
applying a lossless transform to the pixel data prior to compressing the pixel data.
3. The method of claim 1 , wherein the pixel data includes a plurality of pixel quads and compressing the pixel data comprises iteratively compressing each of the plurality of pixel quads.
4. The method of claim 3 , wherein at least two of the plurality of pixel quads are compressed using different processing modes.
5. The method of claim 1 , wherein psychovisually compressing the pixel data comprises removing psychovisually unnecessary bits from the pixel data.
6. The method of claim 1 , where psychovisually compressing the pixel data further comprises:
based on the determined processing mode, calculating average chroma values from the pixel data;
encoding the average chroma values with a first set of bits;
encoding luminance values from the pixel data with a second set of bits; and
inserting the first set of bits and the second set of bits into a compression stream.
7. The method of claim 6 , further comprising:
inserting data in the compression stream that indicates the processing mode used to encode the pixel data.
8. An image processing apparatus, comprising:
an image sensor configured to capture pixel data; and
one or more processors coupled to the image sensor and configured to:
determine a processing mode from a number of available processing modes based on characteristics of the pixel data and one or more simplex boundaries and exclusion in a color subspace;
select a psychovisual compression process based on the determined processing mode; and
psychovisually compress the pixel data from a first data size to a second data size using the selected psychovisual compression process, where the second data size is less than the first data size and the first data size and the second data size have a fixed ratio.
9. The apparatus of claim 8 , further comprising:
applying a lossless transform to the pixel data prior to compressing the pixel data.
10. The apparatus of claim 8 , wherein the pixel data includes a plurality of pixel quads and compressing the pixel data comprises iteratively compressing each of the plurality of pixel quads.
11. The apparatus of claim 10 , wherein at least two of the plurality of pixel quads are compressed using different processing modes.
12. The apparatus of claim 8 , wherein psychovisually compressing the pixel data comprises removing psychovisually unnecessary bits from the pixel data.
13. The apparatus of claim 8 , where psychovisually compressing the pixel data further comprises:
based on the determined processing mode, calculating average chroma values from the pixel data; and
encoding the average chroma values with a first set of bits;
encoding luminance values from the pixel data with a second set of bits; and
inserting the first set of bits and the second set of bits into a compression stream.
14. The apparatus of claim 13 further comprising:
inserting data in the compression stream that indicates the processing mode used to encode the pixel data.