[rothron]

Seems like a slightly unfair comparison. Training the compressor moves data from the images into the compressor, making the bit per pixel evaluation slightly more iffy.

[huhtenberg]

Not really.

As long as the decompressor needs just an image file and no other data, it's a fair game.

[bhouston]

How large is the decompressor to download?

Is this image compression tool good at images it was not trained on?

How bad does it get in those situations?

Is this training data fixed into the codex forever? Will there be slightly different image codexs that have different training data? That would be sort of hellish.

[rothron]

You'd need some bits in the file telling you which trained decoder you need to get the proper image out.

What would the image of that girl look like if they used WaveOne Aerial?

How is this not cheating if the example image they used to compare algorithms is in the training data?

[ClassyJacket]

As long as they don't test on the training set it's a fair comparison isn't it?