Digital image processing is my subject. I've read these filters in my course and seeing their visualization here really made my concept much clearer. Thanks for providing such great stuff.
Great video!! It's so interesting to learn about these complex processes that are behind everyday things like filters etc that we usually take for granted
Thank you, Leo - your videos are fantastic! I also watched your video on image/video compression, and I found it engaging, and really enjoyed it. It's cool to see how much your production quality has improved between then and now! Your video then (~3 years ago) was awesome, but you've really taken your channel to a whole different level since then. Thank you so much for your work, I really appreciate you helping me satisfy my curiosity!
Thanks! Sometimes the algorithm decides to promote videos months, or even years after they are made. I had a few videos like that in the past, this video may become one of those too 🙂
Amazing video, thanks a lot!!! Could you explain or possibly make another video explaining how to move around in the embedding space to edit specific detail of the image? Anyways thanks again.
Thanks! Arxiv Insights has a nice video explaining that: ruclips.net/video/dCKbRCUyop8/видео.html This is not the only way to manipulate properties of images using neural networks though. If we have a dataset with labeled attributes, then we can also use conditional GANs.
@@leoisikdogan oh thanks for pointing out that video... And yes I'm aware of cgans, but my dataset doesn't have labels/attributes... so bad luck 😛. Though info gan would work here... but still moving through the embedding space looks more awesome from my experience :), though finding a specific embedding with specific detail is not easy... but yes with the video you provided it seems quite implementable. Thanks again.
Can you elaborate more on gamma correction in the context of image filtering and computer vision? I’m aware that a neural network can learn to apply gamma correction. But why don’t they gamma expand the input images?
My colleagues and I actually did some research on this. We trained an image preprocessor with a trainable gamma parameter as well as other parameters to improve computer vision performance. You can find the full paper here: www.isikdogan.com/files/isikdogan2019_visionisp_icip.pdf
This is awesome! Still waiting to see the eye contact AI go public. Maybe you could take a look at Snapchat's Lens Studio machine learning implementation (and possibly make a lens). I have made a few lenses using the Lens Studio system and the results turned out well on mobile as well as desktop (they have an application available for Windows). EDIT: Microsoft has included an “Eye Contact” option for the Surface Pro X laptop built into Windows. No other devices are supported.
Thanks! I have no information on the eye contact project since I left Intel :) I'm no longer working on it. I actually did take a look at Snapchat's Lens Studio recently. Their face landmark detection model seems pretty robust!
@@leoisikdogan Thanks for the reply. It seems the only eye contact application currently available is one implemented on Windows by Microsoft for the Surface Pro X. Perhaps one day we will get lucky and see an open-source implementation that utilizes the GPU resources on most enthusiast desktops (rather than a dedicated chip included in a specific laptop model.) I’ve subscribed and will be looking forward to more of your detailed and insightful content. :)
Digital image processing is my subject. I've read these filters in my course and seeing their visualization here really made my concept much clearer. Thanks for providing such great stuff.
This video is so thorough, Leo! Such a great, concise overview!
Can't believe only 203 people like this video. The explanation is great!!!
What a brilliant overview @leo! Eager to see more...
Thank you for the simple and informative step by step explanation. the whole subject was simple and very clear.👏👏👏
your content is so detailed and concise! loved it and wishing for the success of your channel.
Thanks!
Great video!! It's so interesting to learn about these complex processes that are behind everyday things like filters etc that we usually take for granted
Thank you, Leo - your videos are fantastic! I also watched your video on image/video compression, and I found it engaging, and really enjoyed it. It's cool to see how much your production quality has improved between then and now! Your video then (~3 years ago) was awesome, but you've really taken your channel to a whole different level since then.
Thank you so much for your work, I really appreciate you helping me satisfy my curiosity!
Thanks! I'm happy to hear that!
Great video! Nicely covered!
Well explained all videos helps me a lot
This video is so well made... I cant believe it has so little views.
Thanks! Sometimes the algorithm decides to promote videos months, or even years after they are made. I had a few videos like that in the past, this video may become one of those too 🙂
Great Works ! Keep it up !!!!
Amazing video, thanks a lot!!! Could you explain or possibly make another video explaining how to move around in the embedding space to edit specific detail of the image?
Anyways thanks again.
Thanks! Arxiv Insights has a nice video explaining that: ruclips.net/video/dCKbRCUyop8/видео.html
This is not the only way to manipulate properties of images using neural networks though. If we have a dataset with labeled attributes, then we can also use conditional GANs.
@@leoisikdogan oh thanks for pointing out that video... And yes I'm aware of cgans, but my dataset doesn't have labels/attributes... so bad luck 😛. Though info gan would work here... but still moving through the embedding space looks more awesome from my experience :), though finding a specific embedding with specific detail is not easy... but yes with the video you provided it seems quite implementable.
Thanks again.
5:09 I'll remember this forever, was even asked in our midsem exam hahaah
hi thanks for your effort, could you please make a video on deconvolution and transpose convolution?
Great video!
great work
Can you elaborate more on gamma correction in the context of image filtering and computer vision?
I’m aware that a neural network can learn to apply gamma correction. But why don’t they gamma expand the input images?
My colleagues and I actually did some research on this. We trained an image preprocessor with a trainable gamma parameter as well as other parameters to improve computer vision performance. You can find the full paper here: www.isikdogan.com/files/isikdogan2019_visionisp_icip.pdf
This is awesome! Still waiting to see the eye contact AI go public. Maybe you could take a look at Snapchat's Lens Studio machine learning implementation (and possibly make a lens). I have made a few lenses using the Lens Studio system and the results turned out well on mobile as well as desktop (they have an application available for Windows).
EDIT: Microsoft has included an “Eye Contact” option for the Surface Pro X laptop built into Windows. No other devices are supported.
Thanks! I have no information on the eye contact project since I left Intel :) I'm no longer working on it.
I actually did take a look at Snapchat's Lens Studio recently. Their face landmark detection model seems pretty robust!
@@leoisikdogan Thanks for the reply. It seems the only eye contact application currently available is one implemented on Windows by Microsoft for the Surface Pro X. Perhaps one day we will get lucky and see an open-source implementation that utilizes the GPU resources on most enthusiast desktops (rather than a dedicated chip included in a specific laptop model.) I’ve subscribed and will be looking forward to more of your detailed and insightful content. :)
good video~
Great content👌🔥
what is the name of the app ?
Please upload a video on the detailed working of image enhancements using AI
Helal olsun...
2:03 conveloutional
🇹🇷🧿🇹🇷
clickbait !!!! why didnt you show that girl filter ??😠😠😠😠😠😠
Amazing video!