AI Paper Atlas

Research map / Benchmarks / FSC-147

FSC-147 leaderboard

21 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on FSC-147. Best MAE: 5.74 (CountGD*, NeurIPS 2024).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

FSC-147 · MAE

21 papers, each with its best reported number for its own method; lower is better.

11 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

FSC-147 · MAE (val)

17 papers, each with its best reported number for its own method; lower is better.

7 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

FSC-147 · RMSE

14 papers, each with its best reported number for its own method; lower is better.

4 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

FSC-147 · RMSE (val)

10 papers, each with its best reported number for its own method; lower is better.

#RMSE ↓MethodPaperTopic
126.08CountGD*
GroundingDINO
CountGD: Multi-Modal Open-World Counting
NeurIPS 2024 · Table 1
Human and embodied modeling
230.24PBECount
ResNet34 U-Net
PBECount: Prompt-Before-Extract Paradigm for Class-Agnostic Counting
AAAI 2025 · Table 1
CNN features
332.01URM
ResNet50
Single Domain Generalization for Few-Shot Counting via Universal Representation Matching
CVPR 2025 · Table 6
Zero-/few-shot and meta-learning
454.93CountSE (text)
GroundingDINO-B
CountSE: Soft Exemplar Open-set Object Counting
ICCV 2025 · Table 1
Open-set and few-shot detection
555.52HER-Count
Qwen3-VL-2B
HER-Count: Learning Hyper-Exemplar Representation for Generalized Zero-Shot Object Counting
ECCV 2026 · Table 4
CLIP and zero-shot
658.78T2ICount
Stable Diffusion v1.5
T2ICount: Enhancing Cross-modal Understanding for Zero-Shot Counting
CVPR 2025 · Table 1
CLIP and zero-shot
760.78RCACClass-Agnostic Object Counting Robust to Intraclass Diversity
ECCV 2022 · Table 2
Open-set and few-shot detection
869.07FamNet
ResNet-50
[Re] Learning to count everything
NeurIPS 2022 · Table: Comparison with Few-Sho
RNNs and normalization
969.07FamNet
ResNet-50
Learning To Count Everything
CVPR 2021 · Table: Comparison with Few-Sho
Open-set and few-shot detection
1088.63Ours (Patch-Selection)
Ours (Base)
Zero-Shot Object Counting
CVPR 2023 · Table 1
Open-set and few-shot detection

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.