(2018) and a better title that doesn't say CNN, c'mon
Learning to See in the Dark (2018)
71–80 of 179 posts
Re: Learning to See in the Dark (2018)
#72It's surprising how little code [1] is needed to do this. On the other hand I feel this is quite dependent on the specific camera models and might not work on the RAW data downloaded from my phone. Happy to be corrected. [1] - https://github.com/cchen156/Learning-to-See-in-the-Dark/blob...
It's a huge amount of code, hidden behind the tensorflow import statements. It's common to credit GPUs for the rapid spread of deep learning, but good GPUs were available for quite a few years before deep learning really took off. As someone who wrote * a lot* of OpenCL code, including my own python wrappers, I'm fairly certain this code would be thousands of lines without a computation graph framework library. These…
We don't have good GPUs that is as fast until 580 (or to some extents, the first Titan). Previous generations only about 2 to 5x faster depending on what types of CPU you compare against.
IMHO, credit should always go to Alex Krizhevsky for the rapid spread of deep learning. He has shown us it was possible. Even without Tensorflow and PyTorch, we will be fine with Caffe, torch, mxnet or Julia.
Re: Learning to See in the Dark (2018)
#73Earlier quoted context omitted.
The title needs to be changed so brains recognize it as such. It either needs a preceding adjective or letter indicating what type of convolutional network it is. The other option is spelling it out. Most people will read CNN as the news channel. Even those familiar with neural networks.
Not everyone is from the US...
A significant part are though and it would be helpful for that segment.
Re: Learning to See in the Dark (2018)
#74As a photographer, the comparison to "raw" results without color balance or noise removal seems somewhat deceptive. The effects visible in the video seem easy to quickly replicate with existing techniques, such as the "surface blur" filter that averages out pixel values in areas with similar color. This happens at the expense of detail in low-contrast areas, producing a plastic-like appearance of human skin and hair,…
To me the results seem vastly superior to those sort of simple DSP algorithms. The video shows a comparison with some denoising: https://youtu.be/qWKUFK7MWvg?t=102
Re: Learning to See in the Dark (2018)
#75The problem with techniques like this is that they fundamentally amount to ‘making a plausible guess as to what the image would look like’, since essentially they can’t extract information that is simply not there. There is a Shannon entropy limit here. Machine learning is really machine-enhanced educated-guesswork, which has its place but also has its limits.
Agreed. It gets to an important point of the purpose of the photo. Photo as a record versus photo as an aesthetic piece. This hurts the photo as a record but improves the photo as an aesthetic piece. This would be a bad addition to a security camera, but perhaps a good addition to an instagram pipeline. There are plenty of other issues there, like is it good/healthy for stuff like instagram to be diverging away from records, creating unrealistic (or perhaps literally unreal) expectations. We're taking what are slowly inching closer to imaginative art pieces and presenting them as records.
Re: Learning to See in the Dark (2018)
#76Earlier quoted context omitted.
The title needs to be changed so brains recognize it as such. It either needs a preceding adjective or letter indicating what type of convolutional network it is. The other option is spelling it out. Most people will read CNN as the news channel. Even those familiar with neural networks.
Not everyone is from the US...
We get CNN.com outside the US as well.
Re: Learning to See in the Dark (2018)
#77Earlier quoted context omitted.
I, too, was initially very confused by the headline.
Well, Turner used to colorize b/w movies, I guess progress marches on.
Well I got the right CNN by the time I got to the end of the title, but, "Turner"? The British pastoral landscape painter? You lost me.
Re: Learning to See in the Dark (2018)
#78I wonder if photographic evidence "enhanced" by such a method would be admissible in court?
Perhaps it depends on "which court". The legal systems in some parts of the world would allow it.
Re: Learning to See in the Dark (2018)
#79Earlier quoted context omitted.
CNN here is a "Convolutional Neural Network"
The title needs to be changed so brains recognize it as such. It either needs a preceding adjective or letter indicating what type of convolutional network it is. The other option is spelling it out. Most people will read CNN as the news channel. Even those familiar with neural networks.
yeah, i read it exactly as CNN == Cable News Network and was confused for while...
Re: Learning to See in the Dark (2018)
#80Earlier quoted context omitted.
CNN here is a "Convolutional Neural Network"
The title needs to be changed so brains recognize it as such. It either needs a preceding adjective or letter indicating what type of convolutional network it is. The other option is spelling it out. Most people will read CNN as the news channel. Even those familiar with neural networks.
People will have the intellectual curiosity to click in to see what the title means anyways, I don't see how this is a problem in need of a fix.