icdar-2017-mlt.

Dataset from Papers With Code

Submit a result ↵

§ 01 · Leaderboard

Best published scores.

54 results indexed across 4 metrics. Shaded row marks current SOTA; ties broken by submission date.

Primary: accuracy · higher is better
All metrics: f-measure, h-mean, precision, recall

f-measure

17 rows

#	Model	Org	Submitted	Paper / code	f-measure
01	PMTD*	—	Mar 2019	Pyramid Mask Text Detector · code	80.13
02	SBD	—	Dec 2019	Exploring the Capacity of an Orderless Box Discretizatio… · code	79.47
03	TextBPN++ (ResNet-50+DCN)OSS	Zhang et al.	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	77.48
04	CharNet H-88	—	Oct 2019	Convolutional Character Networks · code	75.77
05	DBNet (ResNet-50+DCN)OSS	Liao et al. (USTC)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	74.70
06	GNNets	—	Sep 2019	Geometry Normalization Networks for Accurate Scene Text … · code	74.54
07	PAN	—	Nov 2018	Mask R-CNN with Pyramid Attention Network for Scene Text…	74.30
08	SPCNET	—	Nov 2018	Scene Text Detection with Supervised Pyramid Context Net… · code	74.10
09	CharNet R-50	—	Oct 2019	Convolutional Character Networks · code	73.42
10	PSENet-1s	—	Jun 2018	Shape Robust Text Detection with Progressive Scale Expan… · code	72.45
11	Corner Localization (multi-scale)	—	Feb 2018	Multi-Oriented Scene Text Detection via Corner Localizat… · code	72.40
12	TextBPN++ (ResNet-50)OSS	Zhang et al. (HCIILAB)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	72.33
13	PSENet (ResNet-152)	—	Mar 2019	Shape Robust Text Detection with Progressive Scale Expan… · code	72.13
14	DBNet (ResNet-18+DCN)OSS	Liao et al. (USTC)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	71.70
15	FOTS MS	—	Jan 2018	FOTS: Fast Oriented Text Spotting with a Unified Network · code	70.75
16	FOTS	—	Jan 2018	FOTS: Fast Oriented Text Spotting with a Unified Network · code	67.25
17	Corner Localization (single-scale)	—	Feb 2018	Multi-Oriented Scene Text Detection via Corner Localizat… · code	66.80

h-mean

1 row

#	Model	Org	Submitted	Paper / code	h-mean
01	CRAFT	—	Apr 2019	Character Region Awareness for Text Detection · code	73.90

precision

18 rows

#	Model	Org	Submitted	Paper / code	precision
01	PMTD*	—	Mar 2019	Pyramid Mask Text Detector · code	84.42
02	Corner Localization (single-scale)	—	Feb 2018	Multi-Oriented Scene Text Detection via Corner Localizat… · code	83.80
03	TextBPN++ (ResNet-50+DCN)OSS	Zhang et al.	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	83.74
04	DBNet (ResNet-50+DCN)OSS	Liao et al. (USTC)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	83.10
05	SBD	—	Dec 2019	Exploring the Capacity of an Orderless Box Discretizatio… · code	82.75
06	DBNet (ResNet-18+DCN)OSS	Liao et al. (USTC)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	81.90
07	FOTS MS	—	Jan 2018	FOTS: Fast Oriented Text Spotting with a Unified Network · code	81.86
08	CharNet H-88	—	Oct 2019	Convolutional Character Networks · code	81.27
09	FOTS	—	Jan 2018	FOTS: Fast Oriented Text Spotting with a Unified Network · code	80.95
10	CRAFT	—	Apr 2019	Character Region Awareness for Text Detection · code	80.60
11	SPCNET	—	Nov 2018	Scene Text Detection with Supervised Pyramid Context Net… · code	80.60
12	TextBPN++ (ResNet-50)OSS	Zhang et al. (HCIILAB)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	80.49
13	PAN	—	Nov 2018	Mask R-CNN with Pyramid Attention Network for Scene Text…	80
14	GNNets	—	Sep 2019	Geometry Normalization Networks for Accurate Scene Text … · code	79.63
15	CharNet R-50	—	Oct 2019	Convolutional Character Networks · code	77.07
16	PSENet-1s	—	Jun 2018	Shape Robust Text Detection with Progressive Scale Expan… · code	77.01
17	PSENet (ResNet-152)	—	Mar 2019	Shape Robust Text Detection with Progressive Scale Expan… · code	75.35
18	Corner Localization (multi-scale)	—	Feb 2018	Multi-Oriented Scene Text Detection via Corner Localizat… · code	74.30

recall

18 rows

#	Model	Org	Submitted	Paper / code	recall
01	SBD	—	Dec 2019	Exploring the Capacity of an Orderless Box Discretizatio… · code	76.44
02	PMTD*	—	Mar 2019	Pyramid Mask Text Detector · code	76.25
03	TextBPN++ (ResNet-50+DCN)OSS	Zhang et al.	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	72.10
04	CharNet H-88	—	Oct 2019	Convolutional Character Networks · code	70.97
05	Corner Localization (multi-scale)	—	Feb 2018	Multi-Oriented Scene Text Detection via Corner Localizat… · code	70.60
06	CharNet R-50	—	Oct 2019	Convolutional Character Networks · code	70.10
07	GNNets	—	Sep 2019	Geometry Normalization Networks for Accurate Scene Text … · code	70.06
08	PAN	—	Nov 2018	Mask R-CNN with Pyramid Attention Network for Scene Text…	69.80
09	PSENet (ResNet-152)	—	Mar 2019	Shape Robust Text Detection with Progressive Scale Expan… · code	69.18
10	SPCNET	—	Nov 2018	Scene Text Detection with Supervised Pyramid Context Net… · code	68.60
11	PSENet-1s	—	Jun 2018	Shape Robust Text Detection with Progressive Scale Expan… · code	68.40
12	CRAFT	—	Apr 2019	Character Region Awareness for Text Detection · code	68.20
13	DBNet (ResNet-50+DCN)OSS	Liao et al. (USTC)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	67.90
14	TextBPN++ (ResNet-50)OSS	Zhang et al. (HCIILAB)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	65.67
15	DBNet (ResNet-18+DCN)OSS	Liao et al. (USTC)	May 2022	Arbitrary Shape Text Detection via Boundary Transformer	63.80
16	FOTS MS	—	Jan 2018	FOTS: Fast Oriented Text Spotting with a Unified Network · code	62.30
17	FOTS	—	Jan 2018	FOTS: Fast Oriented Text Spotting with a Unified Network · code	57.51
18	Corner Localization (single-scale)	—	Feb 2018	Multi-Oriented Scene Text Detection via Corner Localizat… · code	55.60

Fig 2 · Rows sorted by score within each metric. Shaded row marks SOTA. Dates reflect model or paper release where available, otherwise the date Codesota accessed the source.

§ 04 · Literature

12 papers
tied to this benchmark.

Every paper below corresponds to at least one row in the leaderboard above. Click through for the arXiv preprint and, when available, the reference implementation.

Arbitrary Shape Text Detection via Boundary Transformer
IEEE Transactions on Multimedia 2023May 2022·TextBPN++ (ResNet-50+DCN), DBNet (ResNet-50+DCN), TextBPN++ (ResNet-50) +1
arXiv ↗
Exploring the Capacity of an Orderless Box Discretization Network for Multi-orientation Scene Text Detection
Dec 2019·SBD
arXiv ↗Code
Convolutional Character Networks
Oct 2019·CharNet H-88, CharNet R-50
arXiv ↗Code
Geometry Normalization Networks for Accurate Scene Text Detection
Sep 2019·GNNets
arXiv ↗Code
Character Region Awareness for Text Detection
Apr 2019·CRAFT
arXiv ↗Code
Pyramid Mask Text Detector
Mar 2019·PMTD*
arXiv ↗Code
Shape Robust Text Detection with Progressive Scale Expansion Network
Mar 2019·PSENet (ResNet-152)
arXiv ↗Code
Mask R-CNN with Pyramid Attention Network for Scene Text Detection
Nov 2018·PAN
arXiv ↗
Scene Text Detection with Supervised Pyramid Context Network
Nov 2018·SPCNET
arXiv ↗Code
Shape Robust Text Detection with Progressive Scale Expansion Network
Jun 2018·PSENet-1s
arXiv ↗Code
Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
Feb 2018·Corner Localization (multi-scale), Corner Localization (single-scale)
arXiv ↗Code
FOTS: Fast Oriented Text Spotting with a Unified Network
Jan 2018·FOTS MS, FOTS
arXiv ↗Code

§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result ↵Read submission guide

What a submission needs

01A public checkpoint or API endpoint
02A reproduction script with frozen commit + seed
03Declared evaluation environment (Python, deps)
04One row per metric declared by this dataset
05A contact so we can follow up on discrepancies

icdar-2017-mlt.

Best published scores.

12 paperstied to this benchmark.

Neighbouring benchmarks.

Have a score that beatsthis table?

12 papers
tied to this benchmark.

Have a score that beats
this table?