sniklaus

5 ms·

1.
Creating Animations Between Two Images(twitter.com)
1 points by sniklaus 5y ago|0 comments
2.
by sniklaus 5y ago
Video demo: https://twitter.com/simon_niklaus/status/1450863810160459777
3.
arXiv DOOM(sniklaus.com)
3 points by sniklaus 5y ago|1 comments
4.
by sniklaus 7y ago
That means that there were 175574 HTTP GET requests, but my guess is that if someone fast-forwards that that might trigger another request. So the actual number of views may be a little lower.
5.
by sniklaus 7y ago
I see your point now. This post originally pointed to the teaser video and not the project website (I do not know who changed that, must have been an admin): http://content.sniklaus.com/kenburns/promo.mp4
6.
by sniklaus 7y ago
That looks awesome, thank you for sharing!
7.
by sniklaus 7y ago
Only very few people in comparison watched the full video. Here are the logs for the content server from yesterday 5 PM PST, covering the previous 24 hours (search for /kenburns/promo.mp4): https://pastebin.com/4zT
8.
by sniklaus 7y ago
Looks like there is another source of traffic at the moment that dwarfs HN and I do not know where it comes from: https://imgur.com/qok0Wmk
9.
by sniklaus 7y ago
Thank you for sharing your thoughts! We designed the automatic camera path estimation to minimize the amount of disocclusion which indeed simplifies the problem. As you correctly pointed out, inpainting the background is an additional chall
10.
by sniklaus 7y ago
Good suggestion, thank you!
11.
by sniklaus 7y ago
It is the result of a lot of work (a lot) and I truly believe everyone in my situation with the right mindset would have been able to achieve it. Being a Ph.D. student and having a chance to do internships at companies like Adobe or Google
12.
by sniklaus 7y ago
I just got the daily email from my server with some metrics. It has experienced over 500 gigabytes of traffic in the last few hours and the video is only 3 megabytes in size. I definitely did not expect such a HN-effect.
13.
by sniklaus 7y ago
It seems like he is using them extensively but not too extreme. Looks great, thank you for sharing!
14.
by sniklaus 7y ago
It's complicated but the gist of it is that they are trying to commercialize it, yes. For what it is worth, you might be able to find it in commercial software once they were successful with their business endeavors. 2. Yes and there a
15.
by sniklaus 7y ago
Those masks look great, thank you for the pointer!
16.
by sniklaus 7y ago
I unfortunately did not get the approval to release the context-aware frame interpolation, but one of my older works is open source for research purposes: https://github.com/sniklaus/pytorch-sepconv 1. I am not sure ab
17.
by sniklaus 7y ago
Since I have done the work while interning at Adobe, I am afraid that I will not be able to commercialize it myself due to patenting. I agree with you that it has great potential, there are still quite a few failure cases though and it is a
18.
by sniklaus 7y ago
There definitely is room for improvement. Our proposed framework uses Mask R-CNN for object segmentation. However, Mask R-CNN only as a resolution of 28x28 pixels for the object masks if I remember correctly (which is then resized to the ac
19.
by sniklaus 7y ago
Yeah, we just are good at seeing artifacts in human renderings, it can easily look wrong or uncanny. In contrast, if a boulder is rendered geometrically wrong we usually do not notice it.
20.
by sniklaus 7y ago
I was thinking about what I should post and decided to make this little teaser since it demonstrates the gist of our work within a few seconds. I understand that the typically Hacker News audience may find the paper more worthwhile though.
21.
by sniklaus 7y ago
I wholeheartedly agree with you and believe that depth perception is a key aspect that is is missing in the status quo of viewing still images!
22.
by sniklaus 7y ago
The effect itself is actually not too uncommon. It is just very tedious and time-consuming to do manually since it requires segmenting the elements in the scene, arranging them in a 3D space, filling in any holes and specifying a virtual ca
23.
by sniklaus 7y ago
More or less. Estimating the depth from a single image is highly challenging and far from solved. We thus had to made sure that the depth estimate is suitable for synthesizing new views. And we are not explicitly modelling layers, we actual
24.
by sniklaus 7y ago
Thank you for sharing your thoughts! * There definitely should be more variety in terms of camera paths. Our framework either supports fully automatic results (as shown) or ones with a manual camera path. For an automatically generated slid
25.
by sniklaus 7y ago
That is called guided upsampling and we are in fact doing just that to get a better estimate of the scene geometry. Specifically, we estimate the scene geometry at a resolution of 256x256 pixels (for a 1:1 aspect ratio, it varies for other
26.
by sniklaus 7y ago
I just watched the trailer to that special, really interesting insights into the vision of sharks. And kudos on the nifty special effects to illustrate it!
27.
by sniklaus 7y ago
That is a nice way of putting it, thank you for chiming in!
28.
by sniklaus 7y ago
I would be happy if it gets its own name in the future. I am voting against "sniklaus" effect though since this was a team effort and I could not have done it without my mentors.
29.
by sniklaus 7y ago
Thank you for clarifying, I see where you are coming from! My background is in research and it is fairly common to extend / reuse an existing terminology, but I totally understand and accept your point of view.
30.
by sniklaus 7y ago
Our work focuses on physically correct depth. We have noticed that such effects created by professional artists emphasize the parallax to an extend that is not physically correct. As such, artists still seem to know best how to animate a ca

More ›