Codesota · Benchmark · msra-td500Home/Leaderboards/Vision & Documents/Scene Text Detection/msra-td500
Unknown

msra-td500.

msra-td500 is a state-of-the-art machine learning benchmark indexed on Codesota. This page tracks published model results, top scores per metric, and the SOTA timeline for msra-td500.

Paper ↗Leaderboard ↓
§ 01 · SOTA history

Year over year.

§ 02 · Leaderboard

Results by metric.

Found a wrong score or missing run?
Use row edits to send a sourced correction into moderation.
Add / edit result ↗Report issue ↗

Fps

Fps is the reported evaluation metric for msra-td500. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Fpsverifiedpapervendorcommunityunverified
RankModelTrustScoreYearLinksFix
01FAST-T-512
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified137.22021Paper ↗Code ↗Looks wrong?
02DBNet++ (ResNet-18) (512)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified802022Paper ↗Code ↗Looks wrong?
03FAST-T-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified79.62021Paper ↗Code ↗Looks wrong?
04FAST-S-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified722021Paper ↗Code ↗Looks wrong?
05FAST-B-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified56.82021Paper ↗Code ↗Looks wrong?
06DBNet++ (ResNet-18) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified552022Paper ↗Code ↗Looks wrong?
07DBNet++ (ResNet-50) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified292022Paper ↗Code ↗Looks wrong?
08MixNet
From paper: MixNet: Toward Accurate Detection of Challenging Scene Text in the Wild
verified15.22023Paper ↗Code ↗Looks wrong?

Precision

Precision is the reported evaluation metric for msra-td500. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Precisionverifiedpapervendorcommunityunverified
RankModelTrustScoreYearLinksFix
01BPDO
ResNet-50+FPN+DCN. Table 1. ICASSP 2024.
verified94.662024Paper ↗Looks wrong?
02TextBPN++ (ResNet-50+DCN)
ResNet-50+DCN backbone. Table VII. T-MM 2023.
verified93.692022Paper ↗Looks wrong?
03TextBPN++ (ResNet-18)
ResNet-18 backbone. Table VII. T-MM 2023.
verified92.382022Paper ↗Looks wrong?
04HTBNet (ResNet-50)
ResNet-50 backbone. Table 9. MDPI Entropy, Jun 2024.
verified92.22024Paper ↗Looks wrong?
05FAST-B-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified92.12021Paper ↗Code ↗Looks wrong?
06RMIPN
Table 2. ICASSP 2024.
verified92.12024Paper ↗Looks wrong?
07FAST-S-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified91.62021Paper ↗Code ↗Looks wrong?
08DBNet++ (ResNet-50) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified91.52022Paper ↗Code ↗Looks wrong?
09DB-ResNet-50 (736)
From paper: Real-time Scene Text Detection with Differentiable Binarization
verified91.52019Paper ↗Code ↗Looks wrong?
10DPNet (ResNet-50, 736px)
ResNet-50, 736×736 input. Table 7. PLOS ONE, Oct 2024.
verified91.432024Paper ↗Looks wrong?
11FAST-T-512
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified91.12021Paper ↗Code ↗Looks wrong?
12MixNet
From paper: MixNet: Toward Accurate Detection of Challenging Scene Text in the Wild
verified90.72023Paper ↗Code ↗Looks wrong?
13DBNet++ (ResNet-18) (512)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified89.72022Paper ↗Code ↗Looks wrong?
14CRAFT
From paper: Character Region Awareness for Text Detection
verified88.22019Paper ↗Code ↗Looks wrong?
15FAST-T-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified88.12021Paper ↗Code ↗Looks wrong?
16DBNet++ (ResNet-18) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified87.92022Paper ↗Code ↗Looks wrong?
17FTSN + MNMS
From paper: Fused Text Segmentation Networks for Multi-oriented Scene Text Detection
verified87.62017Paper ↗Looks wrong?
18Corner Localization
From paper: Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
verified87.62018Paper ↗Code ↗Looks wrong?
19EAST + PVANET2x
From paper: EAST: An Efficient and Accurate Scene Text Detector
verified87.282017Paper ↗Code ↗Looks wrong?
20RRD∗
From paper: Rotation-Sensitive Regression for Oriented Scene Text Detection
verified872018Paper ↗Looks wrong?
21SegLink
From paper: Detecting Oriented Text in Natural Images by Linking Segments
verified862017Paper ↗Code ↗Looks wrong?
22TextSnake
From paper: TextSnake: A Flexible Representation for Detecting Text of Arbitrary Shapes
verified83.22018Paper ↗Code ↗Looks wrong?
23PixelLink+VGG16 2s
From paper: PixelLink: Detecting Scene Text via Instance Segmentation
verified832018Paper ↗Code ↗Looks wrong?

F Measure

F Measure is the reported evaluation metric for msra-td500. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for F Measureverifiedpapervendorcommunityunverified
RankModelTrustScoreYearLinksFix
01BPDO
ResNet-50+FPN+DCN. P=94.66, R=88.48. Table 1. ICASSP 2024. New SOTA on MSRA-TD500.
verified91.472024Paper ↗Looks wrong?
02TextBPN++ (ResNet-50+DCN)
ResNet-50+DCN backbone, 38.5 FPS. P=93.69, R=86.77. Table VII. T-MM 2023.
verified90.12022Paper ↗Looks wrong?
03TextBPN++ (ResNet-18)
ResNet-18 backbone, 38.5 FPS. P=92.38, R=87.46. Table VII. T-MM 2023.
verified89.852022Paper ↗Looks wrong?
04MixNet
From paper: MixNet: Toward Accurate Detection of Challenging Scene Text in the Wild
verified89.42023Paper ↗Code ↗Looks wrong?
05HTBNet (ResNet-50)
ResNet-50 backbone, 30 FPS. P=92.2, R=83.3. Table 9. MDPI Entropy, Jun 2024.
verified87.52024Paper ↗Looks wrong?
06RMIPN
Plug-and-play RMIP module on DB baseline. P=92.1, R=83.4. Table 2. ICASSP 2024.
verified87.52024Paper ↗Looks wrong?
07FAST-B-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified87.32021Paper ↗Code ↗Looks wrong?
08DBNet++ (ResNet-50) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified87.22022Paper ↗Code ↗Looks wrong?
09DPNet (ResNet-50, 736px)
ResNet-50, 736×736 input. P=91.43, R=82.47. Table 7. PLOS ONE, Oct 2024.
verified86.722024Paper ↗Looks wrong?
10FAST-S-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified86.42021Paper ↗Code ↗Looks wrong?
11DBNet++ (ResNet-18) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified85.12022Paper ↗Code ↗Looks wrong?
12FAST-T-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified84.92021Paper ↗Code ↗Looks wrong?
13DB-ResNet-50 (736)
From paper: Real-time Scene Text Detection with Differentiable Binarization
verified84.92019Paper ↗Code ↗Looks wrong?
14FAST-T-512
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified84.52021Paper ↗Code ↗Looks wrong?
15PAN
From paper: Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network
verified84.12019Paper ↗Code ↗Looks wrong?
16CRAFT
From paper: Character Region Awareness for Text Detection
verified82.92019Paper ↗Code ↗Looks wrong?
17DBNet++ (ResNet-18) (512)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified82.62022Paper ↗Code ↗Looks wrong?
18FTSN + MNMS
From paper: Fused Text Segmentation Networks for Multi-oriented Scene Text Detection
verified822017Paper ↗Looks wrong?
19Corner Localization
From paper: Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
verified81.52018Paper ↗Code ↗Looks wrong?
20RRD∗
From paper: Rotation-Sensitive Regression for Oriented Scene Text Detection
verified792018Paper ↗Looks wrong?
21TextSnake
From paper: TextSnake: A Flexible Representation for Detecting Text of Arbitrary Shapes
verified78.32018Paper ↗Code ↗Looks wrong?
22PixelLink+VGG16 2s
From paper: PixelLink: Detecting Scene Text via Instance Segmentation
verified77.82018Paper ↗Code ↗Looks wrong?
23SegLink
From paper: Detecting Oriented Text in Natural Images by Linking Segments
verified772017Paper ↗Code ↗Looks wrong?
24EAST + PVANET2x
From paper: EAST: An Efficient and Accurate Scene Text Detector
verified76.082017Paper ↗Code ↗Looks wrong?

Recall

Recall is the reported evaluation metric for msra-td500. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Recallverifiedpapervendorcommunityunverified
RankModelTrustScoreYearLinksFix
01BPDO
ResNet-50+FPN+DCN. Table 1. ICASSP 2024.
verified88.482024Paper ↗Looks wrong?
02MixNet
From paper: MixNet: Toward Accurate Detection of Challenging Scene Text in the Wild
verified88.12023Paper ↗Code ↗Looks wrong?
03TextBPN++ (ResNet-18)
ResNet-18 backbone. Table VII. T-MM 2023.
verified87.462022Paper ↗Looks wrong?
04TextBPN++ (ResNet-50+DCN)
ResNet-50+DCN backbone. Table VII. T-MM 2023.
verified86.772022Paper ↗Looks wrong?
05PAN
From paper: Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network
verified83.82019Paper ↗Code ↗Looks wrong?
06RMIPN
Table 2. ICASSP 2024.
verified83.42024Paper ↗Looks wrong?
07DBNet++ (ResNet-50) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified83.32022Paper ↗Code ↗Looks wrong?
08HTBNet (ResNet-50)
ResNet-50 backbone. Table 9. MDPI Entropy, Jun 2024.
verified83.32024Paper ↗Looks wrong?
09FAST-B-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified832021Paper ↗Code ↗Looks wrong?
10DBNet++ (ResNet-18) (736)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified82.52022Paper ↗Code ↗Looks wrong?
11DPNet (ResNet-50, 736px)
ResNet-50, 736×736 input. Table 7. PLOS ONE, Oct 2024.
verified82.472024Paper ↗Looks wrong?
12FAST-T-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified81.92021Paper ↗Code ↗Looks wrong?
13FAST-S-736
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified81.72021Paper ↗Code ↗Looks wrong?
14DB-ResNet-50 (736)
From paper: Real-time Scene Text Detection with Differentiable Binarization
verified79.22019Paper ↗Code ↗Looks wrong?
15FAST-T-512
From paper: FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
verified78.82021Paper ↗Code ↗Looks wrong?
16CRAFT
From paper: Character Region Awareness for Text Detection
verified78.22019Paper ↗Code ↗Looks wrong?
17FTSN + MNMS
From paper: Fused Text Segmentation Networks for Multi-oriented Scene Text Detection
verified77.12017Paper ↗Looks wrong?
18DBNet++ (ResNet-18) (512)
From paper: Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
verified76.52022Paper ↗Code ↗Looks wrong?
19Corner Localization
From paper: Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
verified76.22018Paper ↗Code ↗Looks wrong?
20TextSnake
From paper: TextSnake: A Flexible Representation for Detecting Text of Arbitrary Shapes
verified73.92018Paper ↗Code ↗Looks wrong?
21PixelLink+VGG16 2s
From paper: PixelLink: Detecting Scene Text via Instance Segmentation
verified73.22018Paper ↗Code ↗Looks wrong?
22RRD∗
From paper: Rotation-Sensitive Regression for Oriented Scene Text Detection
verified732018Paper ↗Looks wrong?
23SegLink
From paper: Detecting Oriented Text in Natural Images by Linking Segments
verified702017Paper ↗Code ↗Looks wrong?
24EAST + PVANET2x
From paper: EAST: An Efficient and Accurate Scene Text Detector
verified67.432017Paper ↗Code ↗Looks wrong?
§ 04 · Submit a result

Add to the leaderboard.

← Back to Scene Text Detection