Crowds on ShanghaiTech A

Metric: MAE (lower is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	MAE▲	Extra Data	Paper	Date↕	Code
1	EBC-ZIP-B	47.81	No	EBC-ZIP: Improving Blockwise Crowd Counting with...	2025-06-24	Code
2	APGCC	48.8	No	Improving Point-based Crowd Counting and Localiz...	2024-05-17	Code
3	PSL-Net	49.9	No	-	-	-
4	FGENet	51.66	No	FGENet: Fine-Grained Extraction Network for Cong...	2024-01-02	-
5	VMambaCC	51.87	No	VMambaCC: A Visual State Space Model for Crowd C...	2024-05-07	-
6	CLIP-EBC (ViT-B/16)	52.5	No	CLIP-EBC: CLIP Can Count Accurately through Enha...	2024-03-14	Code
7	P2PNet	52.74	No	Rethinking Counting and Localization in Crowds:A...	2021-07-27	Code
8	CLIP-EBC (ResNet50)	54	No	CLIP-EBC: CLIP Can Count Accurately through Enha...	2024-03-14	Code
9	GauNet (ResNet-50)	54.8	No	Rethinking Spatial Invariance of Convolutional N...	2022-06-10	Code
10	LoViTCrowd	54.8	No	-	-	Code
11	M-SFANet+M-SegNet	57.55	No	Encoder-Decoder Based Convolutional Neural Netwo...	2020-03-12	Code
12	SGANet + CL	57.6	No	Crowd Counting via Segmentation Guided Attention...	2019-11-18	Code
13	SGANet	58	No	Crowd Counting via Segmentation Guided Attention...	2019-11-18	Code
14	S-DCNet	58.3	No	From Open Set to Closed Set: Counting Objects by...	2019-08-15	Code
15	SPANet	59.4	No	Learning Spatial Awareness to Improve Crowd Coun...	2019-09-16	-
16	DM-Count	59.7	No	Distribution Matching for Crowd Counting	2020-09-28	Code
17	FusionCount	62.2	No	FusionCount: Efficient Crowd Counting via Multis...	2022-02-28	Code
18	DMCount-EBC	62.3	No	CLIP-EBC: CLIP Can Count Accurately through Enha...	2024-03-14	Code
19	CAN	62.3	No	Context-Aware Crowd Counting	2018-11-26	Code
20	OrdinalEntropy	65.6	No	Improving Deep Regression with Ordinal Entropy	2023-01-21	Code
21	CSRNet-EBC	66.3	No	CLIP-EBC: CLIP Can Count Accurately through Enha...	2024-03-14	Code
22	LSC-CNN	66.4	No	Locate, Size and Count: Accurately Resolving Peo...	2019-06-18	Code
23	SANet	67	No	-	-	Code
24	CSRNet	68.2	No	CSRNet: Dilated Convolutional Neural Networks fo...	2018-02-27	Code
25	ic-CNN	68.5	No	Iterative Crowd Counting	2018-07-26	-
26	IG-CNN	72.5	No	Divide and Grow: Capturing Huge Diversity in Cro...	2018-07-26	-
27	D-ConvNet	73.5	No	-	-	Code
28	CP-CNN	73.6	No	Generating High-Quality Crowd Density Maps using...	2017-08-02	-
29	Liu et al.	73.6	No	Leveraging Unlabeled Data for Crowd Counting by ...	2018-03-08	Code
30	SAFECount	73.7	No	Few-shot Object Counting with Similarity-Aware F...	2022-01-22	Code
31	ACSCP	75.7	No	-	-	Code
32	Switch-CNN	90.4	No	Switching Convolutional Neural Network for Crowd...	2017-08-01	Code
33	Cascaded-MTL	101.3	No	CNN-based Cascaded Multi-task Learning of High-l...	2017-07-30	Code
34	MCNN	110.2	No	-	-	Code
35	Zhang et al.	181.8	No	-	-	-

#1EBC-ZIP-BSOTA
47.81
MAE· 2025-06-24
EBC-ZIP: Improving Blockwise Crowd Counting with Zero-Inflated Poisson Regression Code
#2APGCCSOTA
48.8
MAE· 2024-05-17
Improving Point-based Crowd Counting and Localization Based on Auxiliary Point Guidance Code
#3PSL-Net
49.9
MAE
No paper
#4FGENetSOTA
51.66
MAE· 2024-01-02
FGENet: Fine-Grained Extraction Network for Congested Crowd Counting
#5VMambaCC
51.87
MAE· 2024-05-07
VMambaCC: A Visual State Space Model for Crowd Counting
#6CLIP-EBC (ViT-B/16)
52.5
MAE· 2024-03-14
CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification Code
#7P2PNetSOTA
52.74
MAE· 2021-07-27
Rethinking Counting and Localization in Crowds:A Purely Point-Based Framework Code
#8CLIP-EBC (ResNet50)
54
MAE· 2024-03-14
CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification Code
#9GauNet (ResNet-50)
54.8
MAE· 2022-06-10
Rethinking Spatial Invariance of Convolutional Networks for Object Counting Code
#10LoViTCrowd
54.8
MAE
No paperCode
#11M-SFANet+M-SegNetSOTA
57.55
MAE· 2020-03-12
Encoder-Decoder Based Convolutional Neural Networks with Multi-Scale-Aware Modules for Crowd Counting Code
#12SGANet + CLSOTA
57.6
MAE· 2019-11-18
Crowd Counting via Segmentation Guided Attention Networks and Curriculum Loss Code
#13SGANetSOTA
58
MAE· 2019-11-18
Crowd Counting via Segmentation Guided Attention Networks and Curriculum Loss Code
#14S-DCNetSOTA
58.3
MAE· 2019-08-15
From Open Set to Closed Set: Counting Objects by Spatial Divide-and-Conquer Code
#15SPANet
59.4
MAE· 2019-09-16
Learning Spatial Awareness to Improve Crowd Counting
#16DM-Count
59.7
MAE· 2020-09-28
Distribution Matching for Crowd Counting Code
#17FusionCount
62.2
MAE· 2022-02-28
FusionCount: Efficient Crowd Counting via Multiscale Feature Fusion Code
#18DMCount-EBC
62.3
MAE· 2024-03-14
CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification Code
#19CANSOTA
62.3
MAE· 2018-11-26
Context-Aware Crowd Counting Code
#20OrdinalEntropy
65.6
MAE· 2023-01-21
Improving Deep Regression with Ordinal Entropy Code
#21CSRNet-EBC
66.3
MAE· 2024-03-14
CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification Code
#22LSC-CNN
66.4
MAE· 2019-06-18
Locate, Size and Count: Accurately Resolving People in Dense Crowds via Detection Code
#23SANet
67
MAE
No paperCode
#24CSRNetSOTA
68.2
MAE· 2018-02-27
CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes Code
#25ic-CNN
68.5
MAE· 2018-07-26
Iterative Crowd Counting
#26IG-CNN
72.5
MAE· 2018-07-26
Divide and Grow: Capturing Huge Diversity in Crowd Images with Incrementally Growing CNN
#27D-ConvNet
73.5
MAE
No paperCode
#28CP-CNNSOTA
73.6
MAE· 2017-08-02
Generating High-Quality Crowd Density Maps using Contextual Pyramid CNNs
#29Liu et al.
73.6
MAE· 2018-03-08
Leveraging Unlabeled Data for Crowd Counting by Learning to Rank Code
#30SAFECount
73.7
MAE· 2022-01-22
Few-shot Object Counting with Similarity-Aware Feature Enhancement Code
#31ACSCP
75.7
MAE
No paperCode
#32Switch-CNNSOTA
90.4
MAE· 2017-08-01
Switching Convolutional Neural Network for Crowd Counting Code
#33Cascaded-MTLSOTA
101.3
MAE· 2017-07-30
CNN-based Cascaded Multi-task Learning of High-level Prior and Density Estimation for Crowd Counting Code
#34MCNN
110.2
MAE
No paperCode
#35Zhang et al.
181.8
MAE
No paper