Fixed buffer real-time image compression technique
View Patent ↗Performing real-time compression on an image for target buffer fullness includes dividing the image into N macro-blocks, performing a discrete cosine transformation (DCT) on each of the N macro-blocks, defining a Quantization Parameter Scalar (Q) for each of the N macro-blocks of the image on the DCT being performed, initializing the Quantization Parameter Scalar (Q) for the first Macro-block to a value that correlates to a buffer fullness of a previously compressed image, and monitoring the buffer fullness by comparing the buffer fullness with the target buffer fullness. The N macro-blocks include 16×16 macro-blocks. The Q value is increased to a first new value when the buffer fullness is greater than the target buffer fullness. The Q value is decreased to a second new value when the buffer fullness is less than the target buffer fullness.
1. A method of performing real-time compression on an image for target buffer fullness, said method comprising:
storing an image in a buffer;
dividing said image into N macro-blocks, wherein said N macro-blocks comprises 16×16 macro-blocks;
conducting a compression process of said image by performing a discrete cosine transformation (DCT) on each of said N macro-blocks;
defining a quantization parameter scalar (Q value) for each said N macro-blocks of said image on said DCT being performed;
initializing said quantization parameter scalar (Q value) for the first macro-block to a value that correlates to a buffer fullness of a previously compressed image;
monitoring said buffer fullness by comparing said buffer fullness with said target buffer fullness; and
outputting results of the monitoring step on a display device.
2. The method of claim 1 , further comprising increasing the Q value to a first new value when said buffer fullness is greater than said target buffer fullness, wherein said first new value equals said Q value plus a difference of said comparison of said buffer fullness with said target buffer fullness (Δ).
3. The method of claim 1 , further comprising decreasing the Q value to a second new value when said buffer fullness is less than said target buffer fullness, wherein said second new value equals said Q value minus a difference of said comparison of said buffer fullness with said target buffer fullness (Δ).
4. The method of claim 1 , further comprising biasing a quantization matrix to small values when said buffer fullness is small at a start of said image, wherein said Q values range from 1 to 64 where small values are close to 1.
5. The method of claim 1 , further comprising biasing said quantization matrix to large values when said buffer fullness is substantially full at an end of said image, wherein said Q values range from 1 to 64 where large values are close to 64.
6. The method of claim 1 , wherein said DCT converts said image into a frequency map.
7. The method of claim 1 , wherein an appropriate amount of information is discarded on said DCT being performed, and wherein said compression process divides each DCT output by a quantization coefficient and rounds-off the result to an integer.
8. The method of claim 7 , wherein compression on said image comprises:
performing a zig-zag scan on the output being quantized; and
performing an entropy coding on said quantization coefficient being zig-zag scanned.
9. An apparatus for performing real-time compression on an image for a target buffer fullness, said image comprising 16×16 macro-blocks, said apparatus comprising:
a buffer that stores said image;
a discrete cosine transformation (DCT) unit that performs a DCT on each said 16×16 macro-blocks, wherein said DCT unit converts said image into a frequency map;
a quantization unit that quantizes an output of said DCT, wherein a selected amount of information is discarded, wherein said compression divides each DCT output by a quantization coefficient and rounds-off the result to an integer, wherein said quantization unit comprises:
a control unit that defines a quantization parameter scalar (Q) for each N macro-blocks of said image, wherein said control unit initializes Q to a value that correlates to a buffer fullness of a previously compressed image, wherein said control unit monitors said buffer fullness and compares said buffer fullness with a target buffer fullness, and wherein said control unit sets an appropriate Q value to be used in said quantization unit;
a zig-zag scan unit that performs a zig-zag scan on the output being quantized; and
an entropy coding unit that performs entropy coding on said quantization coefficient being zig-zag scanned.
10. The apparatus of claim 9 , wherein said control unit increases said Q value to a first new value when said buffer fullness is greater than said target buffer fullness, and wherein said first new value equals said Q value plus a difference of said comparison of said buffer fullness with said target buffer fullness (Δ).
11. The apparatus of claim 10 , wherein said control unit decreases said Q value to a second new value when said buffer fullness is less than said target buffer fullness, and wherein said second new value equals said Q value minus a difference of said comparison of said buffer fullness with said target buffer fullness (Δ).
12. The apparatus of claim 10 , wherein said control unit biases a quantization matrix to small values when said buffer fullness is small at a start of said image, and wherein said Q values range from 1 to 64 where small values are close to 1.
13. The apparatus of claim 10 , wherein said control unit biases a quantization matrix to large values when said buffer fullness is substantially full at the end of said image, and wherein said Q values range from 1 to 64 where large values are close to 64.
14. The apparatus of claim 10 , wherein said quantization unit divides a large coefficient of said DCT to represent a data in a least output on said buffer fullness being substantially full.
15. A system for performing real-time compression on an image for a target buffer fullness, said system comprising:
means for dividing said image into N macro-blocks, wherein said N macro-blocks comprises 16×16 macro-blocks;
means for performing a discrete cosine transformation (DCT) on each said N macro-blocks;
means for quantizing each of said N macro-blocks;
means for defining a quantization matrix (Q*[M]) for each said N macro-blocks;
means for initializing Q to a value that correlates to a buffer fullness of a previously compressed image;
means for monitoring said buffer fullness by comparing said buffer fullness with a target buffer fullness;
means for increasing the Q value to a first new value when said buffer fullness is greater than said target buffer fullness;
means for decreasing said Q value to a second new value when said buffer fullness is greater than said target buffer fullness;
means for performing a zig-zag scan for each of said N macro-blocks; and
means for performing entropy encoding for each of said N macro-blocks.
16. The system of claim 15 , further comprising means for biasing said quantization matrix to small values when said buffer fullness is small at a start of said image, wherein said Q values range from 1 to 64 where small values are close to 1.
17. The system of claim 15 , further comprising means for biasing said quantization matrix to large values when said buffer fullness is substantially full at an end of said image, wherein said Q values range from 1 to 64 where large values are close to 64.
18. The system of claim 15 , wherein said DCT converts said image into a frequency map.
19. The system of claim 15 , wherein an appropriate amount of information is discarded on said DCT being performed.
20. The system of claim 19 , further comprising means for dividing each DCT output by a quantization coefficient and rounding-off the result to an integer.