MorphMoe

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 24 on Tapas (warning: JS-heavy site)

Catbox is down so I'll upload the upscaled image later

Source: Instagram

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 24 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 24 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Yuta 2341 | twitter | danbooru

Artist: Risui Ao | pixiv | twitter | danbooru

Full quality: .jpg 2 MB (1211 × 1871)

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 24 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 24 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 24 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

By 出町柳彩都, source

JR Pass is so expensive....

.

.

And the dance is not completed, be patient.

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 23 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Rinotuna | pixiv | twitter | artstation | linktree | patreon | danbooru

Artist: Shycocoa | pixiv | twitter | artstation | danbooru

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 23 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Drawfag | danbooru

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 22 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Dav-19 | deviantart | danbooru

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 22 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 21 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: 泡面之侠 | pixiv

Full quality: .jpg 1 MB (1268 × 1011)

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 21 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original

I also tried Upscayl but that took about 1000x longer and "reinterpreted" the entire picture in an anime style, which made lines thinner, lost detail etc:

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 20 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original
Unlike photos, upscaling digital art with a well-trained algorithm will likely have little to no undesirable effect. Why? Well, the drawing originated as a series of brush strokes, fill areas, gradients etc. which could be represented in a vector format but are instead rendered on a pixel canvas. As long as no feature is smaller than 2 pixels, the Nyquist-Shannon sampling theorem effectively says that the original vector image can therefore be reconstructed losslessly. (This is not a fully accurate explanation, in practice algorithms need more pixels to make a good guess, especially if compression artifacts are present.) Suppose I gave you a low-res image of the flag of South Korea 🇰🇷 and asked you to manually upscale it for printing. Knowing that the flag has no small features so there is no need to guess for detail (this assumption does not hold for photos), you could redraw it with vector shapes that use the same colors and recreate every stroke and arc in the image, and then render them at an arbitrarily high resolution. AI upscalers trained on drawings somewhat imitate this process - not adding detail, just trying to represent the original with more pixels so that it loooks sharp on an HD screen. However, the original images are so low-res that artifacts are basically inevitable, which is why a link to the original is provided.

Artist: Pani | pixiv | artstation | danbooru

Full quality: .png 27 MB (4000 × 6650)

Artist: Onion-Oni aka TenTh from Random-tan Studio
Original post: #Humanization 20 on Tapas (warning: JS-heavy site)

Upscaled by waifu2x (model: upconv_7_anime_style_art_rgb). Original

Reference image is Mandelbrot set zoomed in by a factor of about 1 million, rotated 90° anticlockwise.
𝒙 = -𝓘𝓶(𝒄) = -0.131,825,253,6 ∓ 0.0000011001; 𝒚 = 𝓡𝓮(𝒄) = -0.7436447860 ± 0.0000014668