Computer Vision

Research focused on enabling computers to interpret and understand visual information from images and videos, including tasks such as image classification, object detection, segmentation, and visual recognition.

15 tasks202 datasets94 results

Tasks & Benchmarks

Show all datasets and SOTA results

Object Detection

COCOMicrosoft Common Objects in Context2014

66.12(box-map)ScyllaNet

COCO 2014 valCOCO 2014 Validation Split2014

COCO test-devCOCO test-dev Split2014

COCO val2017COCO 2017 Object Detection (validation split)2014

DIORDIOR (Dataset for Object detection in Optical Remote sensing images)

ImageNet Detection (ILSVRC DET)ImageNet Large Scale Visual Recognition Challenge — Detection (ILSVRC DET)

ImageNet Localization (ILSVRC LOC)ImageNet Large Scale Visual Recognition Challenge — Localization (ILSVRC LOC)

LVIS v1.0Large Vocabulary Instance Segmentation v1.02019

71.4(box-ap)DINO-X

PASCAL VOC 2007PASCAL Visual Object Classes (VOC) Challenge 2007

Pascal VOC 2012Pascal Visual Object Classes Challenge 20122012

80(mAP-coco-pretrain)SSD512 (VGG-16)

Roboflow100-VL (RF100-VL)Roboflow100-VL (RF100-VL)

Image Classification

CIFAR-10Canadian Institute for Advanced Research 102009

99.1(accuracy)DeiT-B Distilled

CIFAR-10CIFAR-10 (Canadian Institute for Advanced Research-10)

CIFAR-100CIFAR-100 (Canadian Institute For Advanced Research)

CIFAR-100Canadian Institute for Advanced Research 1002009

94.55(accuracy)ViT-H/14

CUB (CUB-200-2011)Caltech-UCSD Birds-200-2011 (CUB-200-2011)

DTDDescribable Textures Dataset (DTD)

GEO-Bench (classification suite)GEO-Bench: Toward Foundation Models for Earth Monitoring

Galaxy10Galaxy10 (Galaxy10 / Galaxy10 DECaLS)

ImageNetImageNet (ILSVRC)2009

97.75(top-5-accuracy)SENet

ImageNet RealImageNet ReaL (Reassessed ImageNet Real Labels)

ImageNet V2ImageNet V2 (ImageNetV2)

ImageNet-1KImageNet Large Scale Visual Recognition Challenge 20122012

91(top-1-accuracy)CoCa (finetuned)

ImageNet-1k

ImageNet-RImageNet-R (ImageNet-Rendition)

ImageNet-SImageNet-Sketch (ImageNet-S)

ImageNet-V2ImageNet-V2 Matched Frequency2019

84(top-1-accuracy)Swin Transformer V2 Large

Met (Metropolitan Museum artworks)The Met Dataset (Metropolitan Museum of Art dataset)

ObjectNetObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models

Oxford Flowers-102Oxford 102 Flower Dataset (Oxford Flowers-102)

Places205Places205 (MIT Places Database)

Places365Places365 (Places365-Standard)

Stanford CarsStanford Cars (Cars196)

VTAB (19 tasks)Visual Task Adaptation Benchmark (VTAB)

aircr.FGVCAircraft (FineGrained Visual Classification of Aircraft)

iNat 2017iNaturalist 2017 (iNat 2017) - iNaturalist Species Classification and Detection Dataset

iNat 2018iNaturalist 2018 (iNat 2018) — iNaturalist Species Classification and Detection Dataset

iNat 2019iNaturalist 2019 (iNat Challenge 2019, FGVC6)

iNaturalist 2021iNaturalist 2021 (iNat-2021) Challenge Dataset

Image segmentation

ADE20KADE20K

BRAVO (OOD)BRAVO (BRAVO Semantic Segmentation / BRAVO Challenge dataset)

BSDS500Berkeley Segmentation Dataset (BSDS500)

0.77(ODS)Segment Anything Model (SAM)

COCO 2017 Instance SegmentationMicrosoft COCO 2017 (Instance Segmentation)

46.5(mAP)Segment Anything Model (SAM)

CityScapes

LVIS (Instance Segmentation)

44.7(mAP)Segment Anything Model (SAM)

LoveDALand-cOVEr Domain Adaptation (LoveDA)

Oxford-IIIT PetsOxford-IIIT Pet Dataset

PASCAL VOC 2012PASCAL Visual Object Classes (VOC) 2012

OCR

Fox (English subset, 600-1300 text tokens)Fox — English subset (pages with 600–1300 text tokens)

OCRBenchOCRBench: On the Hidden Mystery of OCR in Large Multimodal Models

860(Score)HunyuanOCR (1B)

OmniDocBench v1.0OmniDocBench v1.02024

OmniDocBench v1.5OmniDocBench v1.52024

olmOCR-BencholmOCR-Bench2025

Image editing

GEdit-BenchGEdit-Bench

ImgEdit (ImgEdit benchmark)ImgEdit: A Unified Image Editing Dataset and Benchmark

KRIS-BenchKRIS-Bench (Knowledge-based Reasoning in Image-editing Systems Benchmark)

PICABenchPICABench: A Comprehensive Benchmark for Physically Realistic Image Editing

RISEBenchRISEBench (Reasoning-Informed viSual Editing Benchmark)

Image generation

CVTG-2KCVTG-2K

GenEvalGenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment

ICE-Bench (Task1-31 Overall)ICE-Bench: A Unified and Comprehensive Benchmark for Image Creating and Editing

ImageNet 1024x1024ImageNet (ILSVRC2012) — 1024×1024 resized variant

ImageNet 256x256ImageNet (ILSVRC2012) 256x256

ImageNet 512x512ImageNet (ILSVRC2012 / ImageNet-1k)

LongText-BenchLongText-Bench

OmniContextOmniContext

OneIG-ENOneIG-Bench: Omni-dimensional Nuanced Evaluation for Image Generation (English track — OneIG-EN)

OneIG-ZHOneIG-Bench — Chinese track (OneIG-Bench-ZH / OneIG-ZH)

TIIF-Bench miniTIIF-Bench (Text-to-Image Instruction Following Benchmark) — mini (compact evaluation subset)

Object counting

CountBench

Open-Vocabulary Object Detection

LVIS (Object Detection)

ODinW13 (subset of ODinW)Object Detection in the Wild (ODinW) — subset: ODinW13

Video classification

COINCOIN: A Large-scale Dataset for Comprehensive Instructional Video Analysis

Diving-48Diving48

Epic-Kitchens-100 (EK100)EPIC-KITCHENS-100 (EK100)

Kinetics-400Kinetics Human Action Video Dataset (Kinetics-400)

Kinetics-4002017

Something-Something V2Something-Something V2

Something-Something V22017

UCF-1012012

UCF101UCF101: A Dataset of 101 Human Action Classes From Videos in the Wild

Video generation

No datasets indexed yet. Contribute on GitHub

3D generation

No datasets indexed yet. Contribute on GitHub

Video segmentation

DAVISDAVIS (Densely Annotated VIdeo Segmentation) / DAVIS 2017

MOSEcoMplex video Object SEgmentation (MOSE)

YouTube-VOSYouTube-VOS: A Large-Scale Video Object Segmentation Benchmark

3D Understanding

CO3Dv2Common Objects in 3D (CO3D) — version 2 (CO3Dv2)

DTUDTU Multi-View Stereo (MVS) Dataset (DTU Robot Image Data Sets)

Re10KRealEstate10K

ScanNet-1500ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes (ScanNet)

Depth estimation

DA-2KDepth Anything 2K Evaluation Benchmark

DDAD (relative)Dense Depth for Autonomous Driving (DDAD)

DIODE (relative)DIODE: A Dense Indoor and Outdoor DEpth Dataset (Relative Depth)

DIODE Outdoor (metric)DIODE: A Dense Indoor and Outdoor DEpth Dataset

ETH3D (relative)ETH3D Multi-View Stereo Benchmark Dataset (Relative Depth)

HyperSim (metric)Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding

KITTI (metric)KITTI Vision Benchmark Suite (Metric Depth)

KITTI (relative)KITTI (relative)

KITTI Depth2012

NYU Depth V22012

NYUv2 (metric)NYU Depth Dataset V2 (Metric Depth)

NYUv2 (relative)NYUv2 (relative)

SUN RGB-D (metric)SUN RGB-D: A RGB-D Scene Understanding Benchmark Suite

ScanNetScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes

Sintel (relative)MPI Sintel Dataset (Depth and Optical Flow Benchmark) (Relative Depth)

Virtual KITTI 2 (metric)Virtual KITTI 2

iBims-1 (metric)iBims-1 (independent Benchmark images and matched scans - version 1)

Few-Shot Image Classification

COCO 2017 CaptionsMicrosoft COCO Captions (COCO 2017 Captions)

COCO 2017 Panoptic SegmentationMicrosoft COCO (Common Objects in Context) — 2017 Panoptic Segmentation

COCO 2017 StuffCOCO-Stuff (COCO 2017 Stuff / COCO-Stuff 164K)

COCO CaptionsCOCO Captions Dataset2015

COCO minivalCOCO minival Split2014

COCO test-challengeCOCO test-challenge Split2014

COCO val2017 (Instance Segmentation)COCO 2017 Object Detection (validation split)

COCO-StuffCommon Objects in COntext-stuff2018

COCO-TextCOCO-Text Dataset2016

COCO-WholeBodyCOCO-WholeBody Dataset2020

CodeForces (CodeElo)CodeElo

ComplexFuncBench

Crossmodal-3600 (XM3600)Crossmodal-3600: A Massively Multilingual Multimodal Evaluation Dataset

DESED

DL3DV-Benchmarks (140)DL3DV Benchmark (140 scenes)

Fluent Speech Commands

Free Music Archive (FMA) Small

GAIA

GAIA2

HELMETHELMET: How to Evaluate Long-context Language Models Effectively and Thoroughly

HammerBench

HiRoomHiRoom

IMC (Image Matching Challenge)Image Matching Challenge — Phototourism (IMC-PT)

ImageNet-HardImageNet-Hard

LOFTLong-Context Frontiers

LVD-142MLVD-142M

LVD-1689MLVD-1689M

Language benchmarks (overall)Language benchmarks (overall)

LibriCount

LibriSpeech-100h

LibriSpeech-Male-Female

MMVPMMVP (Multimodal Visual Patterns) Benchmark

MRCRMulti-turn Response Coherence and Relevance

MTEB

MathVerse

MegaDepth (19)MegaDepth

MuirBench

NSynth-Instruments

NTIRE 2024 Transparent Surface Challenge (relative)NTIRE 2024: HR Depth from Images of Specular and Transparent Surfaces (Booster Dataset) (Relative Depth)

OCRBench v2OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning

SAT-493MSAT-493M (Maxar 493M satellite imagery pretraining dataset)

SWE-Bench

SWE-Bench Pro

SWE-PolyBench

SciVideoBenchSciVideoBench: Benchmarking Scientific Video Reasoning in Large Multimodal Models

ScreenSpot