Quantization in JPEG Image Compression
In JPEG Compression, loss takes place during quantization and during the DCT Transformation.
Why do we get many 0's after quantization after DCT transformation IN JPEG image Compression.
I might have misunderstood your question. Getting "many 0's" is the whole point with quantization. Zero's are coded implicitly by prefixing every symbol in the huffman code with how many zero's til the next coefficient.
Quantization in JPEG compression is accomplished by doing integer division on every coefficient with the corresponding value in the quantization table. If the coefficient is smaller than the value in the quantization table it will become zero.
In the decoder the coefficients are multiplied by the values in the quantization table so that they're restored to their former value (more or less) unless the coefficient is zero.
After DCT transformation, the compressor divides each DCT output value by a "quantization coefficient" and rounds the result to an integer. The larger the quantization coefficient, the more data is lost. Because of the rounding, you may see a lot 0's. The resulting coefficients contain a significant amount of redundant data. The following Huffman compression will losslessly remove the redundancies, resulting in smaller JPEG data.
More info
What @Dragon66 said is correct, the DCT matrix is divided by a quantification matrix and rounded up to an integer, resulting in 0 when the coefficient is high enough.
The reason for this being made, is that human eye is more sensible to noise in lower spatial frequencies, but ignores most of the noise in higher ones.
When you process the pixel matrix with DCT and obtain the resulting matrix, the coefficient in the top left corner represents the average brightness of the pixel block. Moving to the right the coefficients represent increasing horizontal spatial frequency. Moving down, the coefficients represent increasing vertical spatial frequency.
Therefor, the quantification matrix has higher components moving to the right bottom corner. The information lost when a coefficient hits 0 in those positions is less important for the image itself because it represents higher frecuency components.
Full explanation here: http://www.media-matters.net/docs/resources/Digital%20Files/MPEG/MPEG%20Encoding%20Basics.pdf
链接地址: http://www.djcxy.com/p/62886.html上一篇: 为什么DCT变换比视频/图像压缩中的其他变换更受欢迎
下一篇: JPEG图像压缩中的量化