Table 3 Statistics of the object counting datasets, including crowd counting and other fields. Total-, min-, average- and max represent the total number, the minimum, average number and maximum number of instances in the datasets, respectively

LHI\(^1\) [170]

2007

Real-world

–

–

352 \(\times\) 288

–

–

–

–

UCSD [39]

2008

Real-world

2000

800/1200

238 \(\times\) 158

49,885

11

24.9

46

PETS [167]

2010

Real-world

1076

–

384\(\times\)288

18289

0

–

40

Mall [41]

2012

Real-world

2000

800/1200

320 \(\times\) 240

62,325

13

31

53

UCF_CC_50 [38]

2013

Real-world

50

–

2101 \(\times\) 2888

63,974

94

1,280

4,543

AHU-Crowd [171]

2014

Real-world

–

–

–

–

–

–

–

MICC [172]

2014

Real-world

3358

–

–

17630

0

5.25

28

WorldExpo’10 [55]

2015

Real-world

3980

–

576 \(\times\) 720

199,923

1

50.2

253

Indoor\(^1\) [173]

2016

Real-world

570,000

–

704\(\times\)576

–

0

–

59

LHI\(^2\) [174]

2016

Real-world

3,100

–

1280 \(\times\) 720

5,900

–

–

–

AHU-CROWD [175]

2016

Real-world

107

–

–

45,000

58

–

2201

SHT_A [1]

2016

Real-world

482

300/182

589 \(\times\) 868

241,677

33

501.4

3,139

SHT_B [1]

2016

Real-world

716

400/316

768 \(\times\) 1024

88,488

9

123.6

578

Train Station [176]

2017

Real-world

2000

–

256 \(\times\) 256

62581

1

–

53

STF(C5 &C9) [176]

2017

Real-world

788 &600

–

576 \(\times\) 704

–

3

–

65

Shanghai Subway Station [177]

2017

Real-world

3,000

–

–

–

28.78

–

–

Beijing BRT [178]

2018

Real-world

1280

–

640 \(\times\) 360

–

1

–

64

EBP [179]

2018

Real-world

–

–

720 \(\times\) 408

–

–

–

–

Smartcity [9]

2018

Real-world

50

–

1920 \(\times\) 1080

369

1

7.4

14

CrowdFlow [180]

2018

Synthetic

–

–

300\(\sim\)450

–

–

–

–

UCF-QNRF [89]

2018

Real-world

1,535

1201/334

2013 \(\times\) 2902

1,251,642

49

815

12,865

Wildtrack [181]

2018

Real-world

400

360/40

1920 \(\times\) 1080

8,000

–

20

–

CIISR [182]

2019

Real-world

1000

–

1080 \(\times\) 720

–

–

117

–

Venice [99]

2019

Real-world

167

–

1280 \(\times\) 720

–

–

–

–

Indoor\(^2\) [183]

2019

Real-world

148,243

–

352 \(\times\) 288 or 704\(\times\)576

1,834,770

0

12.4

40

CityStreet [155, 169]

2019

Real-world

500

300/200

676 \(\times\) 380

–

70

–

150

ShanghaiTechRGBD [156]

2019

Real-world

2193

1193/1000

1080 \(\times\) 1920

144,512

6

65.9

234

FDST [158]

2019

Real-world

15,000

9000/6000

1920 \(\times\) 1080 and 1280 \(\times\) 720

394,081

9

26.7

57

Crowd Surveillance [159]

2019

Real-world

13,945

–

1342 \(\times\) 840

386,513

–

35

–

ZhengzhouAirport [184]

2019

Real-world

1,111

–

–

49,061

7

–

128

DLR-ACD [161]

2019

Aerial imagery

33

19/14

3619 \(\times\) 5226

226,291

285

6857

24,368

Categorized [185]

2019

Categories counting

553

–

–

16,521

1

29.8

206

GCC [97]

2019

Synthetic

15,212

–

1080 \(\times\) 1920

7,625,843

0

501

3,995

JHU-CROWD [160]

2019

Real-world

4250

3888/1062

1450 \(\times\) 900

1,114,785

–

262

7286

JHU-CROWD++ [59]

2020

Real-world

4372

2722/500

1600 \(\times\) 910

1,515,005

–

346

25,791

NWPU-Crowd [28]

2020

Real-world

5,109

3109/500/1500

2311 \(\times\) 3383

2,133,238

0

418

20,033

DISCO [162]

2020

Real-world

1935

1435/200/300

1920 \(\times\) 1080

170,270

1

87.99

709

MultiviewX [186]

2020

Multi-view

400

360/40

1920 \(\times\) 1080

16,000

–

40

–

MSCC [163]

2021

Real-world

300

170/30/100

–

–

–

–

–

ShanghaiTechRGBD-Syn [157]

2021

Synthetic

16,911

–

1080 \(\times\) 1920

6,245,510

0

369

3,905

RGBT-CC [164]

2021

Real-world

2,030

1030/200/800

640 \(\times\) 480

138,389

–

67.88

–

DroneCrowd [165]

2021

Drone-based

33,600

–

1920 \(\times\) 1080

4,864,280

25

144.8

455

CVCS [166]

2021

Multi-view

28,000

20000/8000

1920 \(\times\) 1080

–

90

–

180

VSCrowd [187]

2022

Video

62,938

49343/13595

1920 \(\times\) 1080

2,344,276

1

37.2

96

Caltech [188]

2012

Pedestrian detection

2000

–

–

15043

6

–

14

TRANCOS [20]

2015

Vehicle counting

1244

403/420/421

640 \(\times\) 480

46,796

9

–

107

Penguins [17]

2016

Penguins counting

80095

–

–

–

0

7.18

67

WIDER FACE [189]

2016

Face detection

32,203

40\(\%\)/10\(\%\)/50\(\%\)

–

393,703

–

–

–

DukeMTMC [168]

2016

tracking, human detection or ReID

over 2 million

–

1920\(\times\)1080

2,700

–

–

–

WebCamT [190]

2017

WebCam traffic counting

60 million

42,200/17800

352\(\times\)240

–

–

–

–

CARPK [191]

2017

Drone view-based car counting

1448

989/459

–

89,777

–

–

–

MTC [192]

2017

Planting counting

361

186/175

–

–

–

–

DCC [144]

2018

Cell counting

177

100/77

–

–

0

34.1

101

Wheat-Spike [193]

2018

wheat spikes counting

20

8/2/10

1k\(\sim\)3k

20,101

749

1005

1287

VisDrone2019 People [194]

2018

Drone-based crowd counting

3347

2392/329/626

969\(\times\)1482

108,464

10

32.41

289

VisDrone2019 Veheicle [194]

2018

Drone-based vehicle counting

5303

3953/364/986

991\(\times\)1511

198,984

10

37.52

349

Sonar Fish [195]

2019

Sonar Fish counting

537

376/41/120

18,032

–

33.58

–

UAVVC [196]

2020

UAV-based vehicle counting

885

500/385

1,024\(\times\)540

32,770

–

–

–

DVSet [197]

2020

Dense vehicle detection

3,894

2,596/1,298

1,455\(\times\)794

149,869

–

38.48

595

FSC-147 [198]

2021

Few-shot counting

6135

3659/1286/1190

774\(\times\)938

343,818

7

56

3701

FSCD-LVIS [199]

2022

Few-shot counting

6195

4000/1181/1014

–

220,670

20

35.76

–

DroneVehile [200]

2022

Dense vehicle detection

56,878

17,990/1,469/8,980

840\(\times\)712

953,087

–

33.51

–

IOCfish5K [201]

2023

indiscernible object counting

56,37

–/–/–

1080\(\times\)1920

659,024

0

117

2,371

MCAC [153]

2024

Multi-class

9283

4756/2413/2114

224\(\times\)224

442,428

1

47.66

300