Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Char-SAM: Turning Segment Anything Model into Scene Text Segmentation Annotator with Character-level Visual Prompts

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Char-SAM shows that a training-free prompt pipeline can turn SAM into a high-quality scene text segmentation annotator, reaching 92.15 F-score on TextSeg.

desk verdict A clever training-free SAM prompting pipeline for scene text segmentation, with a real glyph-prompt idea; but the headline comparison to supervised SOTA is unfair and the multilingual claims rest on an unexamined English-only glyph assumption. read the letter →

arxiv 2412.19917 v1 pith:VMIF77K6 submitted 2024-12-27 cs.CV

Enze Xie , Jiaho Lyu , Daiqing Wu , Huawen Shen , Yu Zhou This is my paper · ORCID
classification cs.CV
keywords scenetextsegmentationSegmentAnythingModelvisualpromptingcharacter-levelboundingboxglyphtemplateannotationgenerationzero-shotSeg
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper's central claim is that SAM can be turned into a scene text segmentation annotator without any training, provided it is prompted at the right granularity. The authors propose Char-SAM, a pipeline that takes existing word-level bounding boxes and transcriptions, refines them into character-level boxes (CBR), and adds glyph-derived positive and negative point prompts (CGR) to cure SAM's over- and under-segmentation on text. On the TextSeg benchmark, the training-free pipeline reaches 84.80 fgIoU and 92.15 F-score, about one point below the supervised state of the art, and it improves a downstream segmentation model's F-score by 0.65 when its generated annotations are added to training. The motivation is to make pixel-level text annotation cheap enough to generate new datasets from existing detection-only datasets such as COCO-Text and MLT17.

What carries the argument

The machinery has two modules. The Character Bounding-box Refinement (CBR) module uses the CRAFT character-aware detector to turn a word-level box into candidate character boxes, then applies a text recognizer and a watershed split on the SAM logical map to separate merged adjacent characters, finally matching character boxes to the characters in the transcription. The Character Glyph Refinement (CGR) module renders each character category across many fonts, thresholds the rendered pixels into foreground and hole pixels, and lets each glyph template vote; pixels whose vote rate exceeds 0.6 become positive point prompts, hole pixels become negative point prompts, and these are fed to SAM together with the character box. The bbox-to-mask capability of SAM is the engine, while the two modules supply the finer character-level visual prompts that make that engine work on text.

What would settle it

Run Char-SAM on a test set of curved or artistic Latin text and non-Latin scripts such as Chinese, Japanese, or Arabic with manual character-level masks, and compare per-script F-score; a large drop on scripts without glyph templates, or on stylized letters whose template votes place wrong prompts, would falsify the fixed-glyph assumption.

Watch

Extended reading notes

Core claim

Char-SAM's central discovery is that SAM's failure on scene text is not a segmentation-capability problem but a prompt-granularity problem. When prompted with a word-level box, SAM treats the whole word as one object; when prompted with a character-level box alone, it over-segments holes (the inside of 'D' or 'A') and under-segments large characters. The paper shows that adding a CBR module to split word boxes into character boxes, and a CGR module that votes over font-rendered glyph templates to place positive points on character strokes and negative points on holes, resolves both failure modes. The resulting pipeline, still using the off-the-shelf SAM-B model, achieves 84.80 fgIoU and 92.15 F-score on TextSeg, comparable to supervised methods evaluated with multi-scale inference. The authors further show that the masks it generates on COCO-Text and MLT17 are cleaner than the weakly supervised COCO TS and MLT S annotations, and that training on these generated masks improves a downstream TextSeg-trained model.

Load-bearing premise

The CGR module assumes that each character category has a fixed glyph shape across fonts, styles, and writing directions, so that pixels from font-rendered templates can act as reliable positive and negative point prompts for real scene text.

Editorial extensions

If this is right

  • Existing detection datasets with only word-level boxes and transcriptions can be upgraded to pixel-level text masks automatically, at zero training cost.
  • The generated annotations on COCO-Text and MLT17 are of higher quality than previous weakly supervised COCO TS and MLT S masks, so downstream segmentation models trained with them gain +0.65 F-score over the TextSeg baseline.
  • Adding more glyph templates improves the result (84.75 to 84.80 fgIoU and 92.01 to 92.15 F-score when going from 40 to 80 templates), so the font-template library is a tunable resource.
  • The per-prompt ablation shows each added prompt type—character box, positive points, negative points—contributes a measurable gain, confirming that prompt granularity is the operative factor in turning SAM into a text annotator.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper leaves implicit that CGR's fixed-glyph premise is script-dependent: for Chinese, Japanese, Arabic, or heavily stylized Latin logos, the per-character template votes would need a script-specific template library, and the paper reports no coverage analysis for the multilingual MLT17 images.
  • A natural transfer would be to any category with a stable canonical structure (digits, icons, anatomical parts), reusing the same refine-and-vote prompt recipe to fix SAM's over- and under-segmentation on small or hollow objects.
  • Because the downstream training gain comes from replacing weak pseudo-labels with Char-SAM masks, one can imagine a bootstrapping loop—train a text segmentation model on the generated masks, refine the masks with the model, and retrain—which the paper does not explore.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper presents Char-SAM, a training-free pipeline that turns word-level scene text annotations (bounding boxes and transcriptions) into character-level segmentation masks by prompting SAM with character boxes produced by the CBR module and with glyph-derived positive/negative point prompts produced by the CGR module. On TextSeg the method reports 84.80 fgIoU and 92.15 F-score, close to supervised SOTA, and the dataset-swap experiment (Table IV) indicates that training on Char-SAM outputs outperforms training on COCO TS / MLT S pseudo-labels and slightly exceeds the TextSeg-trained baseline.

Significance. If the results hold, Char-SAM would be a useful low-cost annotation tool for scene text segmentation: it is training-free, exploits existing detection datasets, and the Table IV dataset-swap experiment directly tests the usefulness of the generated annotations. The paper's strengths are its clear ablations of CBR/CGR and prompt granularity (Tables II and III) and the downstream training comparison. The main weaknesses are the unverified fixed-glyph assumption behind CGR, the protocol mismatch in the SOTA comparison of Table I, and the absence of variance estimates for the small reported differences.

major comments (3)
  1. [II-C, Table II] The CGR module's central premise, stated as 'the English character, regardless of font or text direction, follows the fixed glyph structure determined by its category,' is not valid for common Latin letterforms (single-storey versus double-storey 'a', single-loop versus double-loop 'g') and is undefined for the Arabic, Chinese, Japanese, Korean, and Bangla scripts that appear in MLT17. Table II attributes +3.0 fgIoU and +4.2 F-score to CGR on TextSeg, and Table IV's downstream gain uses MLT S refined, so the claimed dataset-quality improvement rests on the templates being valid for the actual characters. The paper provides only aggregate gains; please supply a per-character, per-font, or per-script breakdown, and either provide a fallback for non-Latin and variable-glyph characters or explicitly narrow the scope of the MLT17 claim.
  2. [Table I] The comparison in Table I is not a head-to-head zero-shot benchmark. Char-SAM is given the dataset's word-level bounding boxes and transcriptions as input (Section II-A), while the supervised methods are evaluated image-only after being trained on TextSeg masks; the reported 'zero-shot' result is therefore an upper-bound annotation-quality estimate under privileged word-level annotations. Please report Char-SAM with automatically detected boxes and recognized transcriptions, or clearly frame Table I as an annotation-pipeline evaluation rather than a zero-shot segmentation comparison.
  3. [Tables II-IV] The paper reports single runs without error bars or significance tests. This matters most in Table IV, where the claimed benefit over the TextSeg baseline is +0.13 fgIoU and +0.65 F-score, and in Table I, where Char-SAM is within about one point of SOTA; these differences could be within run-to-run variation. Please report multiple seeds (or at least variance) and significance tests, or otherwise justify that the differences exceed noise.
minor comments (4)
  1. [II-B] The CBR module description is underspecified: the recognizer used, the 'elaborated bipartite matching' between character boxes and categories, and the watershed initialization are not described in enough detail for reproduction; please add pseudocode or precise references.
  2. [III-B] The CGR vote threshold (0.6) and the number of collected fonts (80) are introduced without justification; Table II varies only the template number (40 vs. 80), so please add a sensitivity analysis for the threshold or discuss how these choices were selected.
  3. [Tables II and III] There is a small numerical inconsistency: Table II row '#2 (TN=40)' reports 84.75 fgIoU / 92.01 F-score, while Table III row 'Pchar + Ppos + Pneg' reports 84.68 / 91.91, even though the two rows appear to describe the same prompting configuration; please clarify the difference or reconcile the numbers.
  4. [III-E] The visualization in Figure 3 is qualitative; please add per-dataset annotation-quality statistics for COCO TS refined and MLT S refined (e.g., agreement with a manually labeled subset), since Table IV measures downstream performance only.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Char-SAM's masks are scored against independent TextSeg ground truth, and its glyph prompts come from external font templates rather than from the evaluation data.

full rationale

The derivation chain in Char-SAM is not circular. The pipeline takes word-level boxes and transcriptions, refines them with CRAFT, a recognizer, SAM, and watershed (CBR), then builds positive/negative point prompts from font-rendered glyph templates (CGR) and feeds boxes plus points to SAM. The final masks are compared with human ground-truth masks on TextSeg (Table I) and with TextRNet+DeepLabV3+ trained on the generated datasets (Table IV). No result is defined in terms of a quantity fitted to the test set: the glyph templates are generated from collected fonts using ImageDraw/ImageFont, not from TextSeg masks; the threshold 0.6 and 80-font template count are fixed hyperparameters with an ablation (#2 vs #3), not fitted outputs renamed as predictions. The strongest potential concern is the CGR premise that "the English character, regardless of font or text direction, follows the fixed glyph structure determined by its category." That is an assumption about typographic invariance; it may be empirically fragile for variable Latin glyph forms or non-Latin scripts in MLT17, and the paper gives no per-script breakdown, but that is a validity and robustness gap, not circularity, because the assumption is an input premise and the claimed improvement from CGR is measured against independent ground truth. Self-citations appear only for background on related scene-text tasks and are not load-bearing for the central claim. The central result therefore stands on external evidence.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The pipeline relies on external pretrained components (SAM, CRAFT, a text recognizer) and on hand-set hyperparameters, but it introduces no new physical or conceptual entities. Glyph templates are generated from existing fonts rather than postulated.

free parameters (2)
  • CGR vote threshold = 0.6
    Hand-set threshold for retaining glyph-template pixels as prompts; its effect is not ablated.
  • Number of glyph template fonts = 80 (40 in ablation)
    More fonts improve results slightly (#2 vs #3 in Table II); chosen by the authors, not derived.
assumptions (5)
  • domain assumption SAM, when given character-level box and point prompts, produces usable text masks
    The whole pipeline entrusts mask generation to SAM; Sec. II-A and II-C.
  • domain assumption CRAFT trained on SynthText transfers to real-world images well enough for character box refinement
    CBR relies on CRAFT's char boxes; the paper acknowledges a domain gap but does not quantify it; Sec. II-B.
  • domain assumption A text recognizer can reliably assign a single character category to each refined box
    CBR uses the recognizer to create the box-to-character correspondence needed by CGR; Sec. II-B.
  • domain assumption English character glyph structure is fixed enough across fonts and directions for template voting
    CGR generates templates from collected fonts and assumes they apply to scene text; Sec. II-C.
  • domain assumption Word-level bounding boxes and transcriptions are available for the target datasets
    These are the inputs to the pipeline; true for TextSeg, COCO-Text, and MLT17, but the method cannot annotate images without them.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Char-SAM: Turning Segment Anything Model into Scene Text Segmentation Annotator with Character-level Visual Prompts." pith.science (2026). https://pith.science/paper/VMIF77K6

@misc{pith2026241219917,
  author       = {Pith},
  title        = {Pith review of: Char-SAM: Turning Segment Anything Model into Scene Text Segmentation Annotator with Character-level Visual Prompts},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VMIF77K6}},
  note         = {Machine review of arXiv:2412.19917}
}
read the original abstract

The recent emergence of the Segment Anything Model (SAM) enables various domain-specific segmentation tasks to be tackled cost-effectively by using bounding boxes as prompts. However, in scene text segmentation, SAM can not achieve desirable performance. The word-level bounding box as prompts is too coarse for characters, while the character-level bounding box as prompts suffers from over-segmentation and under-segmentation issues. In this paper, we propose an automatic annotation pipeline named Char-SAM, that turns SAM into a low-cost segmentation annotator with a Character-level visual prompt. Specifically, leveraging some existing text detection datasets with word-level bounding box annotations, we first generate finer-grained character-level bounding box prompts using the Character Bounding-box Refinement CBR module. Next, we employ glyph information corresponding to text character categories as a new prompt in the Character Glyph Refinement (CGR) module to guide SAM in producing more accurate segmentation masks, addressing issues of over-segmentation and under-segmentation. These modules fully utilize the bbox-to-mask capability of SAM to generate high-quality text segmentation annotations automatically. Extensive experiments on TextSeg validate the effectiveness of Char-SAM. Its training-free nature also enables the generation of high-quality scene text segmentation datasets from real-world datasets like COCO-Text and MLT17.

Figures

Figures reproduced from arXiv: 2412.19917 by the authors.

Figure 1
Figure 1. Failure cases of SAM on TextSeg [1]. The yellow and red rectangles in [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The overall framework of our Char-SAM, which mainly consists of the Character Bbox Refinement (CBR) module, the Character Glyph Refinement (CGR) module and SAM architecture. annotation process predominantly employs semi-supervised algorithms and suffers from fragmented text masks and mis￾classification of non-text pixels. These issues underscore the necessity of developing a scene text segmentation annotation pipeli… view at source ↗
Figure 3
Figure 3. Comparison of annotation quality of different scene text segmentation datasets. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Prompt Engineering in Segment Anything Model: Methodologies, Applications, and Emerging Challenges

    cs.CV 2025-07 conditional novelty 2.0 of 10

    A structured survey of prompt engineering methods for the Segment Anything Model, covering geometric, textual, and multimodal prompts and their applications.

Reference graph

Works this paper leans on

36 extracted references · 28 canonical work pages · cited by 1 Pith paper

  1. [1]

    Rethink- ing text segmentation: A novel dataset and a text-specific refinement approach,

    X. Xu, Z. Zhang, Z. Wang, B. Price, Z. Wang, and H. Shi, “Rethink- ing text segmentation: A novel dataset and a text-specific refinement approach,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 12 045–12 055

    work page 2021

  2. [2]

    Perceiving ambiguity and semantics without recognition: An efficient and effective ambiguous scene text detector,

    Y . Shu, W. Wang, Y . Zhou, S. Liu, A. Zhang, D. Yang, and W. Wang, “Perceiving ambiguity and semantics without recognition: An efficient and effective ambiguous scene text detector,” in Proceedings of the 31st ACM International Conference on Multimedia , 2023, pp. 1851–1862

    work page 2023

  3. [3]

    Self- training for domain adaptive scene text detection,

    Y . Chen, W. Wang, Y . Zhou, F. Yang, D. Yang, and W. Wang, “Self- training for domain adaptive scene text detection,” in 2020 25th Inter- national Conference on Pattern Recognition (ICPR) . IEEE, 2021, pp. 850–857

    work page 2020

  4. [4]

    SEED: Semantics enhanced encoder-decoder framework for scene text recognition,

    Z. Qiao, Y . Zhou, D. Yang, Y . Zhou, and W. Wang, “SEED: Semantics enhanced encoder-decoder framework for scene text recognition,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 13 528–13 537

    work page 2020

  5. [5]

    PIMNet: A parallel, iterative and mimicking network for scene text recognition,

    Z. Qiao, Y . Zhou, J. Wei, W. Wang, Y . Zhang, N. Jiang, H. Wang, and W. Wang, “PIMNet: A parallel, iterative and mimicking network for scene text recognition,” in Proceedings of the 29th ACM International Conference on Multimedia , 2021, pp. 2046–2055

    work page 2021

  6. [6]

    Arbitrary reading order scene text spotter with local semantics guidance,

    J. Lyu, W. Wang, D. Yang, J. Zhong, and Y . Zhou, “Arbitrary reading order scene text spotter with local semantics guidance,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2025

    work page 2025

  7. [7]

    TPSNet: Reverse thinking of thin plate splines for arbitrary shape scene text representation,

    W. Wang, Y . Zhou, J. Lv, D. Wu, G. Zhao, N. Jiang, and W. Wang, “TPSNet: Reverse thinking of thin plate splines for arbitrary shape scene text representation,” in Proceedings of the 30th ACM International Conference on Multimedia , 2022, pp. 5014–5025

    work page 2022

  8. [8]

    TextCtrl: Diffusion-based scene text editing with prior guidance control,

    W. Zeng, Y . Shu, Z. Li, D. Yang, and Y . Zhou, “TextCtrl: Diffusion-based scene text editing with prior guidance control,” Advances in Neural Information Processing Systems , 2024

    work page 2024

Show all 36 references
  1. [9]

    First creating back- grounds then rendering texts: A new paradigm for visual text blending,

    Z. Li, Y . Shu, W. Zeng, D. Yang, and Y . Zhou, “First creating back- grounds then rendering texts: A new paradigm for visual text blending,” ECAI, 2024

    2024

  2. [10]

    Track the answer: Extending TextVQA from image to video with spatio-temporal clues,

    Y . Zhang, G. Zeng, H. Shen, D. Wu, Y . Zhou, and C. Ma, “Track the answer: Extending TextVQA from image to video with spatio-temporal clues,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2025

    2025

  3. [11]

    Beyond OCR+VQA: Towards end-to-end reading and reasoning for robust and accurate TextVQA,

    G. Zeng, Y . Zhang, Y . Zhou, X. Yang, N. Jiang, G. Zhao, W. Wang, and X.-C. Yin, “Beyond OCR+VQA: Towards end-to-end reading and reasoning for robust and accurate TextVQA,” Pattern Recognition, vol. 138, p. 109337, 2023

    2023

  4. [12]

    LDP: Generalizing to multilin- gual visual information extraction by language decoupled pretraining,

    H. Shen, G. Li, J. Zhong, and Y . Zhou, “LDP: Generalizing to multilin- gual visual information extraction by language decoupled pretraining,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2025

    2025

  5. [13]

    Divide rows and conquer cells: Towards structure recognition for large tables

    H. Shen, X. Gao, J. Wei, L. Qiao, Y . Zhou, Q. Li, and Z. Cheng, “Divide rows and conquer cells: Towards structure recognition for large tables.” in IJCAI, 2023, pp. 1369–1377

    2023

  6. [14]

    Scene text segmentation with text-focused Transformers,

    H. Yu, X. Wang, K. Niu, B. Li, and X. Xue, “Scene text segmentation with text-focused Transformers,” in Proceedings of the 31st ACM International Conference on Multimedia , 2023, pp. 2898–2907

    2023

  7. [15]

    COCO TS dataset: Pixel–level annotations based on weak supervision for scene text segmentation,

    S. Bonechi, P. Andreini, M. Bianchini, and F. Scarselli, “COCO TS dataset: Pixel–level annotations based on weak supervision for scene text segmentation,” in International Conference on Artificial Neural Networks. Springer, 2019, pp. 238–250

    2019

  8. [16]

    Pyramid scene parsing network,

    H. Zhao, J. Shi, X. Qi, X. Wang, and J. Jia, “Pyramid scene parsing network,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 2881–2890

    2017

  9. [17]

    Weak supervi- sion for generating pixel–level annotations in scene text segmentation,

    S. Bonechi, M. Bianchini, F. Scarselli, and P. Andreini, “Weak supervi- sion for generating pixel–level annotations in scene text segmentation,” Pattern Recognition Letters, vol. 138, pp. 1–7, 2020

    2020

  10. [18]

    Hi- SAM: Marrying segment anything model for hierarchical text segmenta- tion,

    M. Ye, J. Zhang, J. Liu, C. Liu, B. Yin, C. Liu, B. Du, and D. Tao, “Hi- SAM: Marrying segment anything model for hierarchical text segmenta- tion,” IEEE Transactions on Pattern Analysis and Machine Intelligence , pp. 1–16, 2024

    2024

  11. [19]

    Looking from a higher-level perspective: Attention and recognition enhanced multi-scale scene text segmentation,

    Y . Ren, J. Zhang, B. Chen, X. Zhang, and L. Jin, “Looking from a higher-level perspective: Attention and recognition enhanced multi-scale scene text segmentation,” in Proceedings of the Asian Conference on Computer Vision (ACCV), December 2022, pp. 3138–3154

    2022

  12. [20]

    Scene text eraser,

    T. Nakamura, A. Zhu, K. Yanai, and S. Uchida, “Scene text eraser,” in 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), vol. 1. IEEE, 2017, pp. 832–837

    2017

  13. [21]

    Editing text in the wild,

    L. Wu, C. Zhang, J. Liu, J. Han, J. Liu, E. Ding, and X. Bai, “Editing text in the wild,” in Proceedings of the 27th ACM international conference on multimedia, 2019, pp. 1500–1508

    2019

  14. [22]

    COCO- Text: Dataset and benchmark for text detection and recognition in natural images,

    A. Veit, T. Matera, L. Neumann, J. Matas, and S. Belongie, “COCO- Text: Dataset and benchmark for text detection and recognition in natural images,” arXiv preprint arXiv:1601.07140 , 2016

    2016 arXiv

  15. [23]

    ICDAR 2017 robust reading challenge on multi-lingual scene text detection and script identification-rrc-mlt,

    N. Nayef, F. Yin, I. Bizid, H. Choi, Y . Feng, D. Karatzas, Z. Luo, U. Pal, C. Rigaud, J. Chazalon et al. , “ICDAR 2017 robust reading challenge on multi-lingual scene text detection and script identification-rrc-mlt,” in 2017 14th IAPR International Conference on Document Ana...

    2017

  16. [24]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Loet al., “Segment anything,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4015–4026

    2023

  17. [25]

    Segment anything in medical images,

    J. Ma, Y . He, F. Li, L. Han, C. You, and B. Wang, “Segment anything in medical images,” Nature Communications, vol. 15, no. 1, p. 654, 2024

    2024

  18. [26]

    Medical SAM Adapter: Adapting segment anything model for medical image segmentation,

    J. Wu, R. Fu, H. Fang, Y . Liu, Z. Wang, Y . Xu, Y . Jin, and T. Arbel, “Medical SAM Adapter: Adapting segment anything model for medical image segmentation,” arXiv preprint arXiv:2304.12620 , 2023

    2023 arXiv

  19. [27]

    ClipSAM: CLIP and SAM collaboration for zero-shot anomaly segmentation,

    S. Li, J. Cao, P. Ye, Y . Ding, C. Tu, and T. Chen, “ClipSAM: CLIP and SAM collaboration for zero-shot anomaly segmentation,” arXiv preprint arXiv:2401.12665, 2024

    2024 arXiv

  20. [28]

    Rsprompter: Learning to prompt for remote sensing instance seg- mentation based on visual foundation model,

    K. Chen, C. Liu, H. Chen, H. Zhang, W. Li, Z. Zou, and Z. Shi, “Rsprompter: Learning to prompt for remote sensing instance seg- mentation based on visual foundation model,” IEEE Transactions on Geoscience and Remote Sensing , 2024

    2024

  21. [29]

    Total-Text: A comprehensive dataset for scene text detection and recognition,

    C. K. Ch’ng and C. S. Chan, “Total-Text: A comprehensive dataset for scene text detection and recognition,” in 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR) , vol. 1. IEEE, 2017, pp. 935–942

    2017

  22. [30]

    TextOCR: Towards large-scale end-to-end reasoning for arbitrary- shaped scene text,

    A. Singh, G. Pang, M. Toh, J. Huang, W. Galuba, and T. Hassner, “TextOCR: Towards large-scale end-to-end reasoning for arbitrary- shaped scene text,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 8802–8812

    2021

  23. [31]

    ICDAR 2019 robust reading challenge on arbitrary-shaped text-rrc-art,

    C. K. Chng, Y . Liu, Y . Sun, C. C. Ng, C. Luo, Z. Ni, C. Fang, S. Zhang, J. Han, E. Ding et al., “ICDAR 2019 robust reading challenge on arbitrary-shaped text-rrc-art,” in 2019 International Conference on Document Analysis and Recognition (ICDAR) . IEEE, 2019, pp. 1571– 1576

    2019

  24. [32]

    Character region aware- ness for text detection,

    Y . Baek, B. Lee, D. Han, S. Yun, and H. Lee, “Character region aware- ness for text detection,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2019, pp. 9365–9374

    2019

  25. [33]

    Synthetic data for text localisation in natural images,

    A. Gupta, A. Vedaldi, and A. Zisserman, “Synthetic data for text localisation in natural images,” in IEEE Conference on Computer Vision and Pattern Recognition, 2016

    2016

  26. [34]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929 , 2020

    2010 arXiv

  27. [35]

    Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,

    L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 40, no. 4, pp. 834–848, 2017

    2017

  28. [36]

    Deep high-resolution representation learning for visual recognition,

    J. Wang, K. Sun, T. Cheng, B. Jiang, C. Deng, Y . Zhao, D. Liu, Y . Mu, M. Tan, X. Wang, W. Liu, and B. Xiao, “Deep high-resolution representation learning for visual recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 43, no. 10, pp. 3349– 3364, 2021

    2021

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.