Browse State-of-the-Art › Image Super-Resolution

Image Super-Resolution

783 papers with code · 69 benchmarks · 43 datasets archive 2025-07-28

Computer Vision

Image Super-Resolution is a machine learning task where the goal is to increase the resolution of an image, often by a factor of 4x or more, while maintaining its content and details as much as possible. The end result is a high-resolution version of the original image. This task can be used for various applications such as improving image quality, enhancing visual detail, and increasing the accuracy of computer vision algorithms.

Description from the archive archive 2025-07-28.

Benchmarks archive 2025-07-28

69 leaderboard tables shown for this task, 69 with rows (a “benchmark” on this site is a table with at least one row, as on /sota), ordered by row count. “Best model” is the first row in the archive's own order at snapshot; nothing is re-ranked here and metric direction is not recorded in the archive. PwC's Trend sparklines are not in the archive, so that column is omitted. 10 shown of 69 until expanded.

Syntology 20 newer papers were judged to report on this task's benchmarks so far (Claude Sonnet 4.5, reading their tables: a model's judgement, not a result; 0 placed): Set14 - 4x upscaling (2 judged, 0 placed), Urban100 - 4x upscaling (2 judged, 0 placed), Manga109 - 4x upscaling (1 judged, 0 placed), DIV2K val - 4x upscaling (1 judged, 0 placed), Set5 - 4x upscaling (1 judged, 0 placed), FFHQ 256 x 256 - 4x upscaling (8 judged, 0 placed), ImageNet (11 judged, 0 placed), CelebA (3 judged, 0 placed), Celeb-HQ 4x upscaling (1 judged, 0 placed). Drawn from 8,886 of 18,637 newer papers checked so far (733 more had nothing to place; 41 could not be read).

DatasetBest model (first row in archive order)PaperCodeSyntologyCompare
Set14 - 4x upscaling (104 rows) DRCT-L DRCT: Saving Image Super-resolution away from Information Bottleneck code Syntology official (archive's flag): 1 ran · 4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified Compare
BSD100 - 4x upscaling (71 rows) DRCT-L DRCT: Saving Image Super-resolution away from Information Bottleneck code Syntology official (archive's flag): 1 ran · 4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified Compare
Urban100 - 4x upscaling (65 rows) Hi-IR-L Hierarchical Information Flow for Generalized Efficient Image Restoration — — Compare
Manga109 - 4x upscaling (50 rows) Hi-IR-L Hierarchical Information Flow for Generalized Efficient Image Restoration — — Compare
Set5 - 2x upscaling (41 rows) DRCT-L DRCT: Saving Image Super-resolution away from Information Bottleneck code Syntology official (archive's flag): 1 ran · 4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified Compare
Set14 - 2x upscaling (35 rows) DRCT-L DRCT: Saving Image Super-resolution away from Information Bottleneck code Syntology official (archive's flag): 1 ran · 4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified Compare
Set5 - 3x upscaling (32 rows) HMA† HMANet: Hybrid Multi-Axis Aggregation Network for Image Super-Resolution code — Compare
BSD100 - 2x upscaling (30 rows) WaveMixSR-V2 WaveMixSR-V2: Enhancing Super-resolution with Higher Efficiency code — Compare
Urban100 - 2x upscaling (29 rows) HMA† HMANet: Hybrid Multi-Axis Aggregation Network for Image Super-Resolution code — Compare
Set14 - 3x upscaling (24 rows) Hi-IR-L Hierarchical Information Flow for Generalized Efficient Image Restoration — — Compare
Urban100 - 3x upscaling (22 rows) Hi-IR-L Hierarchical Information Flow for Generalized Efficient Image Restoration — — Compare
BSD100 - 3x upscaling (21 rows) Hi-IR-L Hierarchical Information Flow for Generalized Efficient Image Restoration — — Compare
DIV2K val - 4x upscaling (21 rows) AESOP Auto-Encoded Supervision for Perceptual Image Super-Resolution code Syntology official (archive's flag): 5 ran · 5 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 1 unverified Compare
Manga109 - 2x upscaling (21 rows) Hi-IR-L Hierarchical Information Flow for Generalized Efficient Image Restoration — — Compare
Manga109 - 3x upscaling (17 rows) Hi-IR-L Hierarchical Information Flow for Generalized Efficient Image Restoration — — Compare
Set5 - 4x upscaling (12 rows) HMA† HMANet: Hybrid Multi-Axis Aggregation Network for Image Super-Resolution code — Compare
FFHQ 256 x 256 - 4x upscaling (11 rows) HiFaceGAN HiFaceGAN: Face Renovation via Collaborative Suppression and Replenishment code — Compare
FFHQ 1024 x 1024 - 4x upscaling (9 rows) HiFaceGAN HiFaceGAN: Face Renovation via Collaborative Suppression and Replenishment code — Compare
IXI (9 rows) EDSR+MMHCA Multimodal Multi-Head Convolutional Attention with Various Kernel... code — Compare
FFHQ 512 x 512 - 4x upscaling (8 rows) HiFaceGAN HiFaceGAN: Face Renovation via Collaborative Suppression and Replenishment code — Compare
Set5 - 8x upscaling (8 rows) MFSRCNN — — — Compare
Set14 - 8x upscaling (7 rows) DBPN-RES-MR64-3 Deep Back-Projection Networks for Single Image Super-resolution code — Compare
VggFace2 - 8x upscaling (7 rows) Full-GWAInet Exemplar Guided Face Image Super-Resolution without Facial Landmarks code — Compare
WebFace - 8x upscaling (7 rows) GFRNet Learning Warped Guidance for Blind Face Restoration code — Compare
BSD100 - 8x upscaling (6 rows) DRLN+ Densely Residual Laplacian Super-Resolution code Syntology 4 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified Compare
ImageNet (6 rows) DAVI Diffusion Prior-Based Amortized Variational Inference for Noisy... code Syntology official (archive's flag): 17 ran · 22 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 5 honoured, 0 violated, 2 with no contract checked; 15 where Syntology's instrument failed) · 5 unverified Compare
CelebA (5 rows) DDNM Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model code Syntology community repositories only · 16 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 0 violated, 3 with no contract checked; 12 where Syntology's instrument failed) · 7 unverified Compare
Manga109 - 8x upscaling (5 rows) DBPN-RES-MR64-3 Deep Back-Projection Networks for Single Image Super-resolution code — Compare
Urban100 - 8x upscaling (5 rows) DRLN+ Densely Residual Laplacian Super-Resolution code Syntology 4 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified Compare
CelebA-HQ 128x128 (4 rows) TSRGP Trustworthy Image Super-Resolution via Generative Pseudoinverse code — Compare
DIV8K val - 16x upscaling (3 rows) Ours w/o cycle-loss One-to-many Approach for Improving Super-Resolution code — Compare
General100 - 4x upscaling (3 rows) SROOE Perception-Oriented Single Image Super-Resolution using Optimal... code — Compare
PIRM-test (3 rows) RankSRGAN RankSRGAN: Generative Adversarial Networks with Ranker for Image... code — Compare
2x upscaling (2 rows) ML-CrAIST ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross... code — Compare
3x upscaling (2 rows) ML-CrAIST ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross... code — Compare
4x upscaling (2 rows) ML-CrAIST ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross... code — Compare
B100 - 2x upscaling (2 rows) ML-CrAIST ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross... code — Compare
B100 - 3x upscaling (2 rows) ML-CrAIST ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross... code — Compare
B100 - 4x upscaling (2 rows) ML-CrAIST ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross... code — Compare
BSDS100 - 2x upscaling (2 rows) DBPN-RES-MR64-3 Deep Back-Projection Networks for Single Image Super-resolution code — Compare
CUFED5 - 4x upscaling (2 rows) Extracter-rec EXTRACTER: Efficient Texture Matching with Attention and Gradient... code — Compare
DIV2K val - 8x upscaling (2 rows) FxSR-PD t=0.8 Flexible Style Image Super-Resolution using Conditional Objective code — Compare
General100 - 8x upscaling (2 rows) FxSR-PD t=0.8 Flexible Style Image Super-Resolution using Conditional Objective code — Compare
Sun80 - 4x upscaling (2 rows) Extracter-rec EXTRACTER: Efficient Texture Matching with Attention and Gradient... code — Compare
BSD100 - 16x upscaling (1 row) ABPN Image Super-Resolution via Attention based Back Projection Networks code — Compare
BSD200 - 2x upscaling (1 row) CSRCNN Cascade Convolutional Neural Network for Image Super-Resolution — — Compare
BSDS100 - 4x upscaling (1 row) DBPN-RES-MR64-3 Deep Back-Projection Networks for Single Image Super-resolution code — Compare
BSDS100 - 8x upscaling (1 row) DBPN-RES-MR64-3 Deep Back-Projection Networks for Single Image Super-resolution code — Compare
Celeb-HQ 4x upscaling (1 row) Edge-informed SR Edge-Informed Single Image Super-Resolution code — Compare
Chikusei Dataset (1 row) DIP-HyperKite (ours) Hyperspectral Pansharpening Based on Improved Deep Image Prior and... code — Compare
DIV2K val - 16x upscaling (1 row) ABPN Image Super-Resolution via Attention based Back Projection Networks code — Compare
DIV8K test - 16x upscaling (1 row) RFB-ESRGAN Perceptual Extreme Super Resolution Network with Receptive Field Block code Syntology 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
EPFL NIR-VIS (1 row) RAMS (ours) Multi-image Super Resolution of Remotely Sensed Images using... code — Compare
General-100 - 4x upscaling (1 row) AESOP Auto-Encoded Supervision for Perceptual Image Super-Resolution code Syntology official (archive's flag): 5 ran · 5 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 1 unverified Compare
KITTI 2012 - 2x upscaling (1 row) PASSRnet Learning Parallax Attention for Stereo Image Super-Resolution code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
KITTI 2012 - 4x upscaling (1 row) PASSRnet Learning Parallax Attention for Stereo Image Super-Resolution code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
KITTI 2015 - 2x upscaling (1 row) PASSRnet Learning Parallax Attention for Stereo Image Super-Resolution code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
KITTI 2015 - 4x upscaling (1 row) PASSRnet Learning Parallax Attention for Stereo Image Super-Resolution code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Manga109 - 16x upscaling (1 row) ABPN Image Super-Resolution via Attention based Back Projection Networks code — Compare
Middlebury - 2x upscaling (1 row) PASSRnet Learning Parallax Attention for Stereo Image Super-Resolution code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Middlebury - 4x upscaling (1 row) PASSRnet Learning Parallax Attention for Stereo Image Super-Resolution code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Set14 (1 row) ATD Transcending the Limit of Local Window: Advanced Super-Resolution... code Syntology official (archive's flag): 11 ran · 11 ran (of which 9 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 2 where Syntology's instrument failed) · 4 unverified Compare
Set5 - 5x upscaling (1 row) HyperRes Hypernetwork-Based Adaptive Image Restoration code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified Compare
Set5 - 6x upscaling (1 row) HyperRes Hypernetwork-Based Adaptive Image Restoration code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified Compare
ShipSpotting (1 row) StableShip Ship in Sight: Diffusion Models for Ship-Image Super Resolution code Syntology official (archive's flag): 6 ran · 6 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 6 where Syntology's instrument failed) · 3 unverified Compare
TextZoom (1 row) NCAP NCAP: Scene Text Image Super-Resolution with Non-CAtegorical Prior code — Compare
Urban100 - 16x upscaling (1 row) ABPN Image Super-Resolution via Attention based Back Projection Networks code — Compare
USR-248 - 4x upscaling (1 row) SRDRM-GAN Underwater Image Super-Resolution using Deep Residual Multipliers code — Compare
WLFW (1 row) ArcFace (0.4) EDFace-Celeb-1M: Benchmarking Face Hallucination with a... code — Compare

Syntology column: samples harvested from the paper's repositories and executed on synthesized fixtures; “ran” is not a correctness claim and does not order the table. Each line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the row's paper, the line starts with that repository's state (“official” is the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from. A dash means no Syntology record for that paper, not a recorded non-run. Each “ran” line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

Libraries

Not in the archive: the export carries no per-task library table, so there is nothing to show at snapshot 2025-07-28.

Datasets archive 2025-07-28

43 datasets whose archive record lists this task, ordered by the archive's paper count. 30 shown of 43 until expanded.

Subtasks archive 2025-07-28

5 subtasks in the archive's task tree.

Parent tasks archive 2025-07-28

Papers archive 2025-07-28

archive papers tagged: 1,589 · with a code link: 783 · where Syntology ran a sample: 188 (164 with a run with no instrument failure, 24 where every run was a failure of Syntology's instrument) Syntology

Show: all tagged papersonly where code ran (188 of 1,589 tagged: 164 with a run with no instrument failure, 24 where every run was a failure of Syntology's instrument)

Page 1 of 16: papers 1 to 100 of 1,589, in archive order: by repositories listed in the archive (most first), then newest first, not by stars (the archive holds no stars, so PwC's “Social” and “Latest” sorts cannot be reproduced). Papers that list no repository come after every paper that lists one.

Papers without a page here are shown as plain text. A Syntology line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the figure “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code); hover it for the repositories the samples that ran came from. The first sentence of the abstract is shown for the first 30 papers; every abstract is on its paper's page.

Syntology lines on 53 of the papers shown; no Syntology record for the others (a paper without an arXiv id cannot be joined to the graph, and absence from the graph layer is not a recorded non-run). “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced; each line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections