Home Console Docs News Papers Tutorials Datasets Wiki SOTA LLM Models GPU Leaderboard Events

English

Object Counting On Fsc147

Metrics

MAE(test)

MAE(val)

RMSE(test)

RMSE(val)

Results

Performance results of various models on this benchmark

					Paper Title	Repository
FamNet	22.08	23.75	99.54	69.07	Learning To Count Everything
Omnicount (Open vocabulary, multi-label, without training)	18.63	-	112	-	OmniCount: Multi-label Object Counting with Semantic-Geometric Priors	-
RCC	17.12	17.49	104.53	58.81	Learning to Count Anything: Reference-less Class-agnostic Counting with Weak Supervision
Counting-DETR	16.79	-	123.56	-	Few-shot Object Counting and Detection
CounTX (uses text descriptions instead of visual exemplars)	15.88	17.10	106.29	65.61	Open-world Text-specified Object Counting
LaoNet	15.78	17.11	97.15	56.81	Object Counting: You Only Need to Look at One	-
BMNet+	14.62	15.74	91.83	58.53	Represent, Compare, and Learn: A Similarity-Aware Framework for Class-Agnostic Counting
SAFECount	14.32	15.28	85.54	47.20	Few-shot Object Counting with Similarity-Aware Feature Enhancement
GCA-SUN	14.00	16.06	92.19	53.04	GCA-SUNet: A Gated Context-Aware Swin-UNet for Exemplar-Free Counting	-
SPDCN	13.51	14.59	96.80	49.97	Scale-Prior Deformable Convolution for Exemplar-Guided Class-Agnostic Counting	-
SemAug-SAFECount	12.74	12.59	89.90	44.95	Semantic Generative Augmentations for Few-Shot Counting
CounTR	11.95	13.13	91.23	49.83	CounTR: Transformer-based Generalised Visual Counting
SemAug-CounTR	11.32	12.31	77.50	41.65	Semantic Generative Augmentations for Few-Shot Counting
LOCA	10.79	10.24	56.97	32.56	A Low-Shot Object Counting Network With Iterative Prototype Adaptation
SSD	9.58	9.73	64.13	29.72	Learning Spatial Similarity Distribution for Few-shot Object Counting
CACViT	9.13	10.63	48.96	37.95	Vision Transformer Off-the-Shelf: A Surprising Baseline for Few-Shot Class-Agnostic Counting
DAVE	8.66	8.91	32.36	28.08	DAVE -- A Detect-and-Verify Paradigm for Low-Shot Counting
GeCo	7.91	9.52	54.28	43.00	A Novel Unified Architecture for Low-Shot Counting by Detection and Segmentation
CountGD	5.74	7.1	24.09	26.08	CountGD: Multi-Modal Open-World Counting

0 of 19 row(s) selected.