Is there any similarity between JPEG compression and image compression via PCA? I remember performing the latter during an EECS16 assignment, but I'm curious as to whether prioritizing low frequencies is a common theme to both techniques.
rheask8246
I think they're slightly different in how they transform/represent the data. PCA compression uses dimensionality reduction, which retains the major features in the data. JPEG compression on the other hand utilizes the discrete cosine transform to compress images.
anzeliu
We can use zig-zag reordering and run-length encoding to further compress our image representation. Run-length encoding is a lossless compression technique that converts a list of values into a list of runs of values. This way, we can remove more redundancy in the quantized DCT coefficients.
Is there any similarity between JPEG compression and image compression via PCA? I remember performing the latter during an EECS16 assignment, but I'm curious as to whether prioritizing low frequencies is a common theme to both techniques.
I think they're slightly different in how they transform/represent the data. PCA compression uses dimensionality reduction, which retains the major features in the data. JPEG compression on the other hand utilizes the discrete cosine transform to compress images.
We can use zig-zag reordering and run-length encoding to further compress our image representation. Run-length encoding is a lossless compression technique that converts a list of values into a list of runs of values. This way, we can remove more redundancy in the quantized DCT coefficients.