Posts by zentrocdot
This the image with the applied filter looks like. I should have thought of showing it that way sooner.
#AI #AIimage #Python #filter #masks
I am adding my image filter step by step at Gitlab. It is better to test them than to take a look at the published images here. The loss in transparency is reducing the working effect.
gitlab.com/zentrocdot/i...
#AI #AIimage #Python #filter
I play arround with some effects on badge creation. Save as png and outer area is transparent.
#AI #AIimage #Python #filter
I play arround with some effects on badge creation. Save as png and outer area is transparent.
#AI #AIimage #Python #filter
I play arround with some effects on badge creation. Save as png and outer area is transparent.
#AI #AIimage #Python #filter
I play arround with some effects on badge creation. Save as png and outer area is transparent.
#AI #AIimage #Python #filter
This comes out if one plays with a new filter. Black is in reality full transparent.
#AI #AIimage #Python #filter
Similarity SSIM: 0.6956 (window 8,8), 0.8362 (window 512,512). Works well.
#image #images #Python #similarity #ssim
Back to SSIM. I added the possibility to add padding to the sliding window. One can speed up the calculation with low lost by using padding.
#image #images #Python #similarity #ssim
With AI, you can test ideas like the following sieve effect. It would take me a while to get that working
#AI #AIimage #filter #Python
Something different for a change. Ai wrote for me a filter for applying a jealousy effect to images. I'm always amazed. Really fast, accurate results, fewer HITL.
#AI #AIimage #filter #Python
Applying SSIM based on the References. My simple approach, my image tiling approach and the sliding window approach using [8,8] pixel moved pixel by pixel (CPU vs. GPU).
#image #images #Python #similarity #ssim
If you interested to see, how good AI works in Programming, take a look here.
www.copus.io/work/5b61b97...
#AI #Python #software #program #script #programming
You can see from the lighter stripe in the image that something went wrong during the upscaling. The color profile of the upscaled image is therefore malformed.
#image #images #Python #imagesimilarity #visualization #upscaling
The human is all the time the better analyzer in the first glance. It is easy to see that the first surface bears little resemblance to the second.
#image #images #Python #imagesimilarity #visualization