Comparison gallery of image scaling algorithms - Revision history

OAbot: Open access bot: url-access updated in citation with #oabot.

2025-05-24T15:05:55Z

Open access bot: url-access updated in citation with #oabot.

← Previous revision		Revision as of 15:05, 24 May 2025
Line 108:		Line 108:
	=== [[Image tracing]] ===		=== [[Image tracing]] ===

	[[Image tracing\|Vectorization]] first creates a resolution-independent vector representation of the graphic to be scaled. Then the resolution-independent version is rendered as a raster image at the desired resolution. This technique is used by [[Adobe Illustrator]] Live Trace, [[Inkscape]], and several recent papers.<ref name="pixelart">{{cite journal\|author=Johannes Kopf and Dani Lischinski\|title=Depixelizing Pixel Art\|journal=ACM Transactions on Graphics \|year=2011\|volume=30 \|issue=4\|pages=99:1–99:8\|doi=10.1145/2010324.1964994\|archiveurl=https://web.archive.org/web/20150901034643/http://research.microsoft.com/en-us/um/people/kopf/pixelart\| archivedate=2015-09-01 \|url=http://research.microsoft.com/en-us/um/people/kopf/pixelart/\|accessdate=24 October 2012}}</ref>		[[Image tracing\|Vectorization]] first creates a resolution-independent vector representation of the graphic to be scaled. Then the resolution-independent version is rendered as a raster image at the desired resolution. This technique is used by [[Adobe Illustrator]] Live Trace, [[Inkscape]], and several recent papers.<ref name="pixelart">{{cite journal\|author=Johannes Kopf and Dani Lischinski\|title=Depixelizing Pixel Art\|journal=ACM Transactions on Graphics \|year=2011\|volume=30 \|issue=4\|pages=99:1–99:8\|doi=10.1145/2010324.1964994\|archiveurl=https://web.archive.org/web/20150901034643/http://research.microsoft.com/en-us/um/people/kopf/pixelart\| archivedate=2015-09-01 \|url=http://research.microsoft.com/en-us/um/people/kopf/pixelart/\|accessdate=24 October 2012\|url-access=subscription}}</ref>
	[[Scalable Vector Graphics]] are well suited to simple geometric images, while photographs do not fare well with vectorization due to their complexity.		[[Scalable Vector Graphics]] are well suited to simple geometric images, while photographs do not fare well with vectorization due to their complexity.

StainedGlassCavern: /* Pixel art scaling algorithms (hqx) */

2025-01-22T16:55:51Z

Pixel art scaling algorithms (hqx)

← Previous revision		Revision as of 16:55, 22 January 2025
Line 70:		Line 70:
	=== [[Pixel art scaling algorithms]] (hqx) ===		=== [[Pixel art scaling algorithms]] (hqx) ===

	For magnifying computer graphics with low resolution and few colors (usually from 2 to 256 colors), better results will be achieved by pixel art scaling algorithms such as [[hqx]] or xbr. These produce sharp edges and maintain high level of detail. Unfortunately due to the standardized size of 218x80 pixels, the "Wiki" image cannot use HQ4x or 4xBRZ to better demonstrate the artifacts they may produce such as row shifting.		For magnifying computer graphics with low resolution and few colors (usually from 2 to 256 colors), better results will be achieved by pixel art scaling algorithms such as [[hqx (algorithm)\|hqx]] or xbr. These produce sharp edges and maintain high level of detail. Unfortunately due to the standardized size of 218x80 pixels, the "Wiki" image cannot use HQ4x or 4xBRZ to better demonstrate the artifacts they may produce such as row shifting.


Line 79:		Line 79:
	\|[[File:Image_after_scaling_(2xBRZ).png\|Image after scaling (2xBRZ)]]		\|[[File:Image_after_scaling_(2xBRZ).png\|Image after scaling (2xBRZ)]]
	\|		\|

	=== [[Pixel art scaling algorithms]] (xbr) ===		=== [[Pixel art scaling algorithms]] (xbr) ===
	The xbr family is very useful for creating smooth edges. It will however deform the shape significantly, which in many cases creates a very appealing result. However it will create an effect similar to [[posterization]] by grouping together local areas into a single colour. It will also remove small details if in-between larger ones which connect together.		The xbr family is very useful for creating smooth edges. It will however deform the shape significantly, which in many cases creates a very appealing result. However it will create an effect similar to [[posterization]] by grouping together local areas into a single colour. It will also remove small details if in-between larger ones which connect together.

Chris the speller: punct

2024-06-14T04:53:07Z

punct

← Previous revision		Revision as of 04:53, 14 June 2024
Line 96:		Line 96:
	An adaptable technique which can deliver variable amounts of detail or smoothness. It aims to preserve the shape and coordinates of original details, without blurring those details into neighboring ones. It will avoid blending pixels which directly touch each other, and instead only blend pixels with their diagonal neighbors.		An adaptable technique which can deliver variable amounts of detail or smoothness. It aims to preserve the shape and coordinates of original details, without blurring those details into neighboring ones. It will avoid blending pixels which directly touch each other, and instead only blend pixels with their diagonal neighbors.

	The "Cutter" name comes from its tendency to cut corners of squares and turn them into diamonds, as well as create distinct faces along stair-stepped pixels, ~~ie:~~ those which exist on along the angles of edges found on a diamond. The "Gem" prefix both refers to the diamond cut, and also many traditional gem cuts which involve cutting corners at a 45 degree angle.		The "Cutter" name comes from its tendency to cut corners of squares and turn them into diamonds, as well as create distinct faces along stair-stepped pixels, i.e. those which exist on along the angles of edges found on a diamond. The "Gem" prefix both refers to the diamond cut, and also many traditional gem cuts which involve cutting corners at a 45-degree angle.

Ellievoid0x: Fixed typo

2024-02-22T18:39:26Z

Fixed typo

← Previous revision		Revision as of 18:39, 22 February 2024
Line 96:		Line 96:
	An adaptable technique which can deliver variable amounts of detail or smoothness. It aims to preserve the shape and coordinates of original details, without blurring those details into neighboring ones. It will avoid blending pixels which directly touch each other, and instead only blend pixels with their diagonal neighbors.		An adaptable technique which can deliver variable amounts of detail or smoothness. It aims to preserve the shape and coordinates of original details, without blurring those details into neighboring ones. It will avoid blending pixels which directly touch each other, and instead only blend pixels with their diagonal neighbors.

	The "Cutter" name comes from its tendency to cut corners of squares and turn them into diamonds, as well as create distinct faces along stair-stepped pixels, ie: those which exist on along the angles of edges found on a diamond. The "Gem" prefix both ~~reffers~~ to the diamond cut, and also many traditional gem cuts which involve cutting corners at a 45 degree angle.		The "Cutter" name comes from its tendency to cut corners of squares and turn them into diamonds, as well as create distinct faces along stair-stepped pixels, ie: those which exist on along the angles of edges found on a diamond. The "Gem" prefix both refers to the diamond cut, and also many traditional gem cuts which involve cutting corners at a 45 degree angle.

AnomieBOT: Dating maintenance tags: {{Cn}}

2023-02-05T19:14:30Z

Dating maintenance tags: {{Cn}}

← Previous revision		Revision as of 19:14, 5 February 2023
Line 61:		Line 61:
	=== [[Image_scaling#Edge-directed_interpolation\|Edge-directed interpolation]] ===		=== [[Image_scaling#Edge-directed_interpolation\|Edge-directed interpolation]] ===

	Edge-directed interpolation algorithms aim to preserve edges in the image after scaling, unlike other algorithms which can produce staircase artifacts around diagonal lines or curves. Examples of algorithms for this task include [[New Edge-Directed Interpolation]] (NEDI),<ref>{{cite web\|url=http://chiranjivi.tripod.com/EDITut.html\|title=Edge-Directed Interpolation\|accessdate=19 February 2016}}</ref><ref>{{cite journal\|author1=Xin Li\|author2=Michael T. Orchard\|journal=2000 IEEE International Conference on Image Processing\|page=311\|url=http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|title=NEW EDGE DIRECTED INTERPOLATION\|access-date=2016-07-03\|archive-url=https://web.archive.org/web/20160214121631/http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|archive-date=2016-02-14\|url-status=dead}}</ref> [[Edge-Guided Image Interpolation]] (EGGI),<ref>{{cite journal \|author1=Zhang, D. \|author2=Xiaolin Wu \|title=An Edge-Guided Image Interpolation Algorithm via Directional Filtering and Data Fusion\|journal=IEEE Transactions on Image Processing \|volume=15 \|issue=8 \|pages=2226–38 \|pmid=16900678 \|bibcode=2006ITIP...15.2226Z \|year=2006 \|doi=10.1109/TIP.2006.877407 \|s2cid=9760560 }}</ref> [[Iterative Curvature-Based Interpolation]] (ICBI),{{cn}} and [[Directional Cubic Convolution Interpolation]] (DCCI).<ref>{{cite web\|author1=Dengwen Zhou\|author2=Xiaoliu Shen\|title=Image Zooming Using Directional Cubic Convolution Interpolation\|url=http://www.mathworks.com/matlabcentral/fileexchange/38570-image-zooming-using-directional-cubic-convolution-interpolation\|accessdate=13 September 2015}}</ref> A study found that DCCI had the best scores in [[Peak signal-to-noise ratio\|PSNR]] and [[Structural similarity\|SSIM]] on a series of test images.<ref>{{cite arXiv\|author1=Shaode Yu\|author2=Rongmao Li\|author3=Rui Zhang\|author4=Mou An\|author5=Shibin Wu\|author6=Yaoqin Xie\|title=Performance evaluation of edge-directed interpolation methods for noise-free images\|eprint=1303.6455\|class=cs.CV\|year=2013}}</ref>		Edge-directed interpolation algorithms aim to preserve edges in the image after scaling, unlike other algorithms which can produce staircase artifacts around diagonal lines or curves. Examples of algorithms for this task include [[New Edge-Directed Interpolation]] (NEDI),<ref>{{cite web\|url=http://chiranjivi.tripod.com/EDITut.html\|title=Edge-Directed Interpolation\|accessdate=19 February 2016}}</ref><ref>{{cite journal\|author1=Xin Li\|author2=Michael T. Orchard\|journal=2000 IEEE International Conference on Image Processing\|page=311\|url=http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|title=NEW EDGE DIRECTED INTERPOLATION\|access-date=2016-07-03\|archive-url=https://web.archive.org/web/20160214121631/http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|archive-date=2016-02-14\|url-status=dead}}</ref> [[Edge-Guided Image Interpolation]] (EGGI),<ref>{{cite journal \|author1=Zhang, D. \|author2=Xiaolin Wu \|title=An Edge-Guided Image Interpolation Algorithm via Directional Filtering and Data Fusion\|journal=IEEE Transactions on Image Processing \|volume=15 \|issue=8 \|pages=2226–38 \|pmid=16900678 \|bibcode=2006ITIP...15.2226Z \|year=2006 \|doi=10.1109/TIP.2006.877407 \|s2cid=9760560 }}</ref> [[Iterative Curvature-Based Interpolation]] (ICBI),{{cn\|date=February 2023}} and [[Directional Cubic Convolution Interpolation]] (DCCI).<ref>{{cite web\|author1=Dengwen Zhou\|author2=Xiaoliu Shen\|title=Image Zooming Using Directional Cubic Convolution Interpolation\|url=http://www.mathworks.com/matlabcentral/fileexchange/38570-image-zooming-using-directional-cubic-convolution-interpolation\|accessdate=13 September 2015}}</ref> A study found that DCCI had the best scores in [[Peak signal-to-noise ratio\|PSNR]] and [[Structural similarity\|SSIM]] on a series of test images.<ref>{{cite arXiv\|author1=Shaode Yu\|author2=Rongmao Li\|author3=Rui Zhang\|author4=Mou An\|author5=Shibin Wu\|author6=Yaoqin Xie\|title=Performance evaluation of edge-directed interpolation methods for noise-free images\|eprint=1303.6455\|class=cs.CV\|year=2013}}</ref>
	\|-		\|-
	\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]		\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]

Citation bot: Add: date. | Use this bot. Report bugs. | Suggested by Headbomb | #UCB_toolbar

2023-02-05T18:54:11Z

Add: date. | Use this bot. Report bugs. | Suggested by Headbomb | #UCB_toolbar

← Previous revision		Revision as of 18:54, 5 February 2023
Line 128:		Line 128:
	=== Deep [[convolutional neural networks]] using perceptual [[Loss function\|loss]] ===		=== Deep [[convolutional neural networks]] using perceptual [[Loss function\|loss]] ===

	Developed on the basis of the super-resolution [[generative adversarial network]] (SRGAN) method,<ref>{{cite web\|url=https://medium.com/@peaceful0907/generative-adversarial-network-and-super-resolution-gan-srgan-e2e7a4440066\|title=Generative Adversarial Network and Super Resolution GAN(SRGAN)\|accessdate=26 August 2020}}</ref> enhanced SRGAN (ESRGAN)<ref>{{cite arXiv\|title=ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks\|eprint=1809.00219\|last1=Wang\|first1=Xintao\|last2=Yu\|first2=Ke\|last3=Wu\|first3=Shixiang\|last4=Gu\|first4=Jinjin\|last5=Liu\|first5=Yihao\|last6=Dong\|first6=Chao\|author7=Chen Change Loy\|last8=Qiao\|first8=Yu\|last9=Tang\|first9=Xiaoou\|year=2018\|class=cs.CV}}</ref> is an incremental tweaking of the same generative adversarial network basis. Both methods rely on a perceptual [[loss function]]<ref>{{cite web\|url=https://deepai.org/machine-learning-glossary-and-terms/perceptual-loss-function\|title=Perceptual Loss Functions\|accessdate=26 August 2020}}</ref> to evaluate training iterations.		Developed on the basis of the super-resolution [[generative adversarial network]] (SRGAN) method,<ref>{{cite web\|url=https://medium.com/@peaceful0907/generative-adversarial-network-and-super-resolution-gan-srgan-e2e7a4440066\|title=Generative Adversarial Network and Super Resolution GAN(SRGAN)\|date=26 April 2020 \|accessdate=26 August 2020}}</ref> enhanced SRGAN (ESRGAN)<ref>{{cite arXiv\|title=ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks\|eprint=1809.00219\|last1=Wang\|first1=Xintao\|last2=Yu\|first2=Ke\|last3=Wu\|first3=Shixiang\|last4=Gu\|first4=Jinjin\|last5=Liu\|first5=Yihao\|last6=Dong\|first6=Chao\|author7=Chen Change Loy\|last8=Qiao\|first8=Yu\|last9=Tang\|first9=Xiaoou\|year=2018\|class=cs.CV}}</ref> is an incremental tweaking of the same generative adversarial network basis. Both methods rely on a perceptual [[loss function]]<ref>{{cite web\|url=https://deepai.org/machine-learning-glossary-and-terms/perceptual-loss-function\|title=Perceptual Loss Functions\|date=17 May 2019 \|accessdate=26 August 2020}}</ref> to evaluate training iterations.
	\|}		\|}

Headbomb: -predatory source

2023-02-05T18:53:51Z

-predatory source

← Previous revision		Revision as of 18:53, 5 February 2023
Line 61:		Line 61:
	=== [[Image_scaling#Edge-directed_interpolation\|Edge-directed interpolation]] ===		=== [[Image_scaling#Edge-directed_interpolation\|Edge-directed interpolation]] ===

	Edge-directed interpolation algorithms aim to preserve edges in the image after scaling, unlike other algorithms which can produce staircase artifacts around diagonal lines or curves. Examples of algorithms for this task include [[New Edge-Directed Interpolation]] (NEDI),<ref>{{cite web\|url=http://chiranjivi.tripod.com/EDITut.html\|title=Edge-Directed Interpolation\|accessdate=19 February 2016}}</ref><ref>{{cite journal\|author1=Xin Li\|author2=Michael T. Orchard\|journal=2000 IEEE International Conference on Image Processing\|page=311\|url=http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|title=NEW EDGE DIRECTED INTERPOLATION\|access-date=2016-07-03\|archive-url=https://web.archive.org/web/20160214121631/http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|archive-date=2016-02-14\|url-status=dead}}</ref> [[Edge-Guided Image Interpolation]] (EGGI),<ref>{{cite journal \|author1=Zhang, D. \|author2=Xiaolin Wu \|title=An Edge-Guided Image Interpolation Algorithm via Directional Filtering and Data Fusion\|journal=IEEE Transactions on Image Processing \|volume=15 \|issue=8 \|pages=2226–38 \|pmid=16900678 \|bibcode=2006ITIP...15.2226Z \|year=2006 \|doi=10.1109/TIP.2006.877407 \|s2cid=9760560 }}</ref> [[Iterative Curvature-Based Interpolation]] (ICBI),~~<ref>~~{{cite journal\|author1=K.Sreedhar Reddy\|author2=Dr.K.Rama Linga Reddy\|journal=International Journal of Advanced Research in Computer and Communication Engineering\|date=December 2013\|volume=2\|issue=12\|page=4631\|url=http://www.ijarcce.com/upload/2013/december/IJARCCE4D-s-sreedhar_reddy_enlargement_of.pdf\|title=Enlargement of Image Based Upon Interpolation Techniques}}~~</ref>~~ and [[Directional Cubic Convolution Interpolation]] (DCCI).<ref>{{cite web\|author1=Dengwen Zhou\|author2=Xiaoliu Shen\|title=Image Zooming Using Directional Cubic Convolution Interpolation\|url=http://www.mathworks.com/matlabcentral/fileexchange/38570-image-zooming-using-directional-cubic-convolution-interpolation\|accessdate=13 September 2015}}</ref> A study found that DCCI had the best scores in [[Peak signal-to-noise ratio\|PSNR]] and [[Structural similarity\|SSIM]] on a series of test images.<ref>{{cite arXiv\|author1=Shaode Yu\|author2=Rongmao Li\|author3=Rui Zhang\|author4=Mou An\|author5=Shibin Wu\|author6=Yaoqin Xie\|title=Performance evaluation of edge-directed interpolation methods for noise-free images\|eprint=1303.6455\|class=cs.CV\|year=2013}}</ref>		Edge-directed interpolation algorithms aim to preserve edges in the image after scaling, unlike other algorithms which can produce staircase artifacts around diagonal lines or curves. Examples of algorithms for this task include [[New Edge-Directed Interpolation]] (NEDI),<ref>{{cite web\|url=http://chiranjivi.tripod.com/EDITut.html\|title=Edge-Directed Interpolation\|accessdate=19 February 2016}}</ref><ref>{{cite journal\|author1=Xin Li\|author2=Michael T. Orchard\|journal=2000 IEEE International Conference on Image Processing\|page=311\|url=http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|title=NEW EDGE DIRECTED INTERPOLATION\|access-date=2016-07-03\|archive-url=https://web.archive.org/web/20160214121631/http://www.csee.wvu.edu/~xinl/papers/ICIP2000a.pdf\|archive-date=2016-02-14\|url-status=dead}}</ref> [[Edge-Guided Image Interpolation]] (EGGI),<ref>{{cite journal \|author1=Zhang, D. \|author2=Xiaolin Wu \|title=An Edge-Guided Image Interpolation Algorithm via Directional Filtering and Data Fusion\|journal=IEEE Transactions on Image Processing \|volume=15 \|issue=8 \|pages=2226–38 \|pmid=16900678 \|bibcode=2006ITIP...15.2226Z \|year=2006 \|doi=10.1109/TIP.2006.877407 \|s2cid=9760560 }}</ref> [[Iterative Curvature-Based Interpolation]] (ICBI),{{cn}} and [[Directional Cubic Convolution Interpolation]] (DCCI).<ref>{{cite web\|author1=Dengwen Zhou\|author2=Xiaoliu Shen\|title=Image Zooming Using Directional Cubic Convolution Interpolation\|url=http://www.mathworks.com/matlabcentral/fileexchange/38570-image-zooming-using-directional-cubic-convolution-interpolation\|accessdate=13 September 2015}}</ref> A study found that DCCI had the best scores in [[Peak signal-to-noise ratio\|PSNR]] and [[Structural similarity\|SSIM]] on a series of test images.<ref>{{cite arXiv\|author1=Shaode Yu\|author2=Rongmao Li\|author3=Rui Zhang\|author4=Mou An\|author5=Shibin Wu\|author6=Yaoqin Xie\|title=Performance evaluation of edge-directed interpolation methods for noise-free images\|eprint=1303.6455\|class=cs.CV\|year=2013}}</ref>
	\|-		\|-
	\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]		\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]

Elliott-AtomicInfinity: This IS bicubic interpolation

2023-02-02T06:49:44Z

This IS bicubic interpolation

← Previous revision		Revision as of 06:49, 2 February 2023
Line 44:		Line 44:
	=== [[Bicubic interpolation]] ===		=== [[Bicubic interpolation]] ===

	Better scaling methods include ~~[[bicubic interpolation]],~~ [[Lanczos resampling]] and [[Mitchell-Netravali filters]].		Better scaling methods include [[Lanczos resampling]] and [[Mitchell-Netravali filters]].
	\|-		\|-
	\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]		\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]
Line 50:		Line 50:
	\|[[File:Image-after-fourier saturated.png\|Fourier-based Interpolation + saturation]]		\|[[File:Image-after-fourier saturated.png\|Fourier-based Interpolation + saturation]]
	\|		\|

	=== [[Fourier transform\|Fourier]]-based interpolation ===		=== [[Fourier transform\|Fourier]]-based interpolation ===

Wikihistorian: added Category:Computing comparisons using HotCat

2023-01-28T05:53:30Z

added Category:Computing comparisons using HotCat

← Previous revision		Revision as of 05:53, 28 January 2023
Line 137:		Line 137:
	[[Category:Image processing]]		[[Category:Image processing]]
	[[Category:Image galleries]]		[[Category:Image galleries]]
			[[Category:Computing comparisons\|Image scaling algorithms]]

Xii-Nyth: finally completed that table that was left unfinished since the creator never did a year later

2022-10-23T07:07:52Z

finally completed that table that was left unfinished since the creator never did a year later

← Previous revision		Revision as of 07:07, 23 October 2022
Line 114:		Line 114:
	\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]		\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]
	\|[[File:Green_Sea_Shell_(EDSR).png\|EDSR]]		\|[[File:Green_Sea_Shell_(EDSR).png\|EDSR]]
	\|[[File:Wiki_(waifu2x).png\|waifu2x]]		\|[[File:Wiki_(waifu2x).png\|waifu2x\|alt=waifu2x (unknown version?)]]
	\|		\|
	=== Deep [[convolutional neural networks]] ===		=== Deep [[convolutional neural networks]] ===
Line 122:		Line 122:
	\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]		\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]
	\|[[File:Green_Sea_Shell_(ESRGAN).png\|ESRGAN]]		\|[[File:Green_Sea_Shell_(ESRGAN).png\|ESRGAN]]
			\|[[File:RealESRGAN-x4plus Time Test Augmentation.png\|alt=RealESRGAN-x4plus TTA]]
	\|
	\|		\|

← Previous revision		Revision as of 07:07, 23 October 2022
Line 114:		Line 114:
	\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]		\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]
	\|[[File:Green_Sea_Shell_(EDSR).png\|EDSR]]		\|[[File:Green_Sea_Shell_(EDSR).png\|EDSR]]
	\|[[File:Wiki_(waifu2x).png\|waifu2x]]		\|[[File:Wiki_(waifu2x).png\|waifu2x\|alt=waifu2x (unknown version?)]]
	\|		\|
	=== Deep [[convolutional neural networks]] ===		=== Deep [[convolutional neural networks]] ===
Line 122:		Line 122:
	\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]		\|[[File:160_by_160_thumbnail_of_%27Green_Sea_Shell%27.png\|160×160 thumbnail reference]]
	\|[[File:Green_Sea_Shell_(ESRGAN).png\|ESRGAN]]		\|[[File:Green_Sea_Shell_(ESRGAN).png\|ESRGAN]]
			\|[[File:RealESRGAN-x4plus Time Test Augmentation.png\|alt=RealESRGAN-x4plus TTA]]
	\|
	\|		\|