Method, apparatus, and system for pre-compression assessment of compressed data length
A method and apparatus are disclosed for image processing using dynamic assessment of image line data to select a compression technique for compressing the image line data prior to storage in a line buffer used for an image processing operation. Compressing the image line of data prior to storage in the line buffer permits reduction in the size of the line buffer. The compressed image data may be decompressed prior to the image processing.
1 . A method of processing image data, comprising:
evaluating a first line of uncompressed image data;
identifying, based on said evaluating of said first line of uncompressed image data, a first compression technique which will compress the first line of uncompressed image data to fit within a line buffer having a bit length less than the bit length of the line of uncompressed image data;
compressing the first line of uncompressed image data using the first compression technique to produce a first line of compressed image data having a first compressed data bit length; and
outputting the first line of compressed image data to the line buffer.
2 . The method of claim 1 , wherein the step of identifying the first compression technique comprises:
predicting, by examining the line of uncompressed image data, a plurality of compressed data bit lengths that would result from compressing the first line of uncompressed image data in accordance with a respective plurality of compression techniques; and
selecting one of said compression techniques which produces a compressed data bit length fitting within the line buffer as the first compression technique.
3 . The method of claim 2 , wherein said step of selecting comprises:
determining, based on the predicted plurality of compressed data bit lengths, a plurality of viable compression techniques that would each produce respectively compressed lines of image data fitting within the line buffer;
determining, from among said plurality of viable compression techniques, a single viable compression technique predicted to produce the greatest compressed data bit length; and
selecting said single viable compression technique as said first compression technique.
4 . The method of claim 1 , further comprising:
sequentially storing the first line of compressed image data into a plurality of FIFO arranged individual line buffers;
reading out the first line of compressed image data from an individual line buffer to a respective decompressor; and
decompressing the first line of compressed image data, read out from the individual line buffer, to produce a first line of decompressed image data.
5 . The method of claim 4 , further comprising image processing using the decompressed first line of image data.
6 . The method of claim 1 , further comprising:
compressing, based on a second selected compression technique different than the first compression technique, a second line of uncompressed image data to produce a second line of compressed image data having a second compressed data bit length different than the first compressed data bit length, the first and second lines of uncompressed image data representing respective first and second image lines forming an image.
7 . The method of claim 6 , further comprising:
storing the first and second lines of compressed image data in FIFO connected line buffers; and
decompressing in parallel the first and second lines of compressed image data stored in the line buffers.
8 . The method of claim 6 , wherein the first and second compression techniques are lossy compression techniques employing different levels of quantization.
9 . The method of claim 6 , wherein the first and second compression technique employ different entropy encoding tables.
10 . The method of claim 9 , wherein the first and second compression techniques are lossless compression techniques.
11 . The method of claim 9 , wherein one of the first and second compression techniques is a lossless compression technique and the other is a lossy compression technique.
12 . The method of claim 9 , wherein the first and second compression techniques are lossy compression techniques employing different levels of quantization.
13 . The method of claim 1 , wherein the steps of compressing and reading out the compressed image data are performed by a pipeline processing circuit.
14 . The method of claim 4 , wherein the image processing performs at a least one of a lens shading correction, color interpolation, a resize operation, color correction, gamma correction, and color conversion.
15 . The method of claim 1 , further comprising:
appending compression information data to said first uncompressed line of image data, said compression information data indicating said first compression technique.
16 . The method of claim 1 , further compressing:
transmitting a compression information signal to a compressor,
wherein said compression information signal indicates the first compression technique and said compressor uses said compression signal to select a compression technique and perform said step of compressing the first line of uncompressed image data.
17 . An imager device comprising:
a pixel cell array configured to output pixel signals in lines of image data;
an analog-to-digital converter configured to convert said pixel signals into pixel data; and
a compression and storage system comprising:
a compressor configured to generate a plurality of lines of compressed image data by sequentially compressing, one by one, a respective plurality of lines of uncompressed image data;
a plurality of individual line buffers in a FIFO arrangement configured to sequentially store, one by one, said lines of compressed image data and to read out, simultaneously, each of said stored lines of compressed image data, each of said plurality of line buffers having a bit length less than a bit length of each of said lines of uncompressed image data; and
an assessor configured to identify, based on an evaluation of each of said plurality of lines of uncompressed image data, a respective compression technique which will compress a respective line of uncompressed image data to generate a respective line of compressed image data fitting within each of said plural individual line buffers, and configured to indicate said identified respective compression to said compressor,
said compressor being configured to compress said respective lines of uncompressed image data using a respective identified compression technique to produce lines of compressed image data.
18 . The imager device of claim 17 , wherein said assessor is further configured to:
predict, by examining a respective line of uncompressed image data, a plurality of compressed data bit lengths that would result from compressing the respective line of uncompressed image data in accordance with a respective plurality of compression techniques; and
select a compression technique for said respective line of uncompressed image data based on said predicted plurality of compressed data bit lengths.
19 . The imager device of claim 18 , wherein said assessor is further configured to:
determine for said respective line of uncompressed image data, based on the predicted plurality of compressed data bit lengths, a plurality of viable compression techniques that would each produce respectively compressed lines of image data fitting within the individual line buffers;
determine, from among said plurality of viable compression techniques, a single viable compression technique predicted to produce the greatest compressed data bit length; and
select said single viable compression technique as a compression technique for said respective lines of uncompressed image data.
20 . The imager device of claim 17 , further comprising:
a decompressor configured to decompress the respective lines of compressed image data read out from the plurality of line buffers.
21 . The imager device of claim 17 , further comprising:
a plurality of decompressors configured to simultaneously receive and decompress said lines of compressed image data from said plurality of line buffers; and
an imaging processing pipeline configured to perform an image line processing operation on said lines of decompressed image data.
22 . The imager device of claim 17 , wherein said assessor is configured to select different compression techniques for different lines of having a second uncompressed image data.
23 . The imager device of claim 22 , wherein said the different compression techniques are lossy compression techniques employing different levels of quantization.
24 . The imager device of claim 22 , wherein said different compression techniques employ different entropy encoding tables.
25 . The imager device of claim 24 , wherein said different compression techniques are lossless compression techniques.
26 . The imager device of claim 24 , wherein one of said different compression techniques is a lossless compression technique and the other is a lossy compression technique.
27 . The imager device of claim 24 , wherein said different compression techniques are lossy compression techniques employing different levels of quantization.
28 . The imager device of claim 17 , wherein said assessor is configured to append compression information data to said uncompressed lines of image data, said compression information data indicating a compression technique for compressing said uncompressed lines of image data.
29 . The imager device of claim 17 , wherein said assessor is configured to transmit a compression information signal to a compressor, and said compression information signal identifies a selected compression technique and said compressor uses said compression information signal to perform said step of compressing a line of uncompressed image data.
30 . The method of claim 4 , further comprising:
storing said first line of uncompressed image data in another line buffer prior to said step of compressing said first line of uncompressed image data;
reading out multiple lines of image data to a processing stage for multiple line processing, said multiple lines of image data including a line of uncompressed image data read out from said another line buffer and lines of decompressed image data read out from respective decompressors of said plurality of FIFO arranged individual line buffers.
31 . The imager device of claim 21 , further comprising:
another line buffer for sequentially storing and reading out said lines of uncompressed image data to said compressor;
a processing stage for processing multiple lines of image data, said multiples lines of image data including a line of uncompressed image data read out from said another line buffer and said lines of decompressed image data read out from said plurality of decompressors.