HiFiC
AI
generative image compression
SHOW HN2020
281points
89comments
#149of 3,028 in 2020
Top 5%of 2020 by points
How HN took it
DebatedReal support and real criticism
18 positive, 27 neutral or asking, 1 mixed, 16 critical, out of 62 comments about the project. Thread score +0.03.
Idea: 10 for, 3 againstPerformance: 3 for, 6 againstExecution: 2 for, 5 againstDocs: 0 for, 2 againstDesign: 1 for, 0 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Results beat JPEG and BPG by a wide margin at low bitrates
From 2 comments
This is amazing. Not only it completely blows JPEG and BPG out of the water in efficiency, but it makes some images look better than the original! It seems to reduce local…
ricardobeat on HN>Currently, this runs at approximately 0.7 megapixel/sec on a GPU with unoptimized TensorFlow. Forgive me, is that in encoding or decoding? And size of library required to decode…
ksec on HN
Noise and skin texture preserved well
From 1 comment
Thanks! Noise is actually preserved really well, and is one of the strengths of using a GAN. Check out this visual example from the "All Evaluation Images" link:…
atorodius on HN
Objections
Fine details like small text and faces are regenerated rather than real
From 3 comments
Yeah, it seems like the HiFiC version has the same large-scale structure, but the details are sort of "regenerated": e.g. it's as if the image format says "there's some hair…
fiddlerwoaroof on HNThis is impressive; but isn't this also susceptible to the type of bug that Xerox copiers got hit with many years ago?…
jasonjayr on HNThanks for the kind words! We hope that the algorithm we presented is a step in the right direction, and we acknowledge that there's more work to be done! Just like with any…
atorodius on HN
Model size (726MB) and slow speed make practical use hard
From 2 comments
Thanks for the kind words! > What size library would be required to decode these type of images? The model is 726MB. Keep in mind that this is a research project - further…
atorodius on HNHi, author here! This is the demo of a research project demonstrating the potential of using GANs for compression. For this to be used in practice, more research is required to…
atorodius on HN
Requests
Compare against AVIF and JPEG XL
From 2 comments
Have you compared the result with latest image compression formats, such as AVIF[0]? [0]: https://aomediacodec.github.io/av1-avif/
keyi on HNI would add JPEG-XL in addition if you're looking for suggestions for other codecs to compare to. It's very competitive with AV1 and beats it, in my opinion, at higher bitrates.…
bscphil on HN
Extend to video with temporal consistency
From 3 comments
I've been waiting for something to implement this concept for so long, and I'm so happy to finally get a chance to explore how it works in practice! It's really fascinating to…
crazygringo on HNSuper exciting! Fantastic work and cool demo! I hope this can be 'stabilized' to allow video compression. If you run eg. style transfer networks on video, the artifacts in the…
isoprophlex on HNHas this been extended to video (beyond compressing each frame separately)?
lokl on HN
Questions
How the adversarial loss works
From 1 comment
> The adversarial loss "learns" what is a compressed image and tries to make the decoder go away from such outputs. Could you expand on this point
nbardy on HN
Training data and compute needed to replicate
From 1 comment
Amazing work! It's something we knew would come at some point, but I didn't expect it to be that good so early! Since some of the authors are in this thread answering questions,…
littlestymaar on HN
Alternatives named
JPEG
From 1 comment
The obvious question when looking at the comparison is, what kind of jpeg did they pick, and was it fair? It turns out that mozjpeg, which is pretty much state of the art, can…
stereo on HN
BPG / HEVC
From 1 comment
>Currently, this runs at approximately 0.7 megapixel/sec on a GPU with unoptimized TensorFlow. Forgive me, is that in encoding or decoding? And size of library required to decode…
ksec on HN
JPEG XL
From 1 comment
I would add JPEG-XL in addition if you're looking for suggestions for other codecs to compare to. It's very competitive with AV1 and beats it, in my opinion, at higher bitrates.…
bscphil on HN
AVIF
From 1 comment
Have you compared the result with latest image compression formats, such as AVIF[0]? [0]: https://aomediacodec.github.io/av1-avif/
keyi on HN