Pith. sign in

REVIEW 4 major objections 4 minor 41 references

CapHDR2IR: Caption-Driven Transfer from Visible Light to Infrared Domain

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper claims that combining high-dynamic-range visible input with dense caption features produces state-of-the-art infrared image generation on the HDRT dataset.

desk verdict A well-ablated HDR+caption RGB-to-IR method whose SOTA claim is conditional on adding three missing baselines and more rigorous evaluation. read the letter →

arxiv 2411.16327 v1 pith:AS3MKX6H submitted 2024-11-25 cs.CV

Jingchao Peng , Thomas Bashford-Rogers , Zhuang Shao , Haitao Zhao , Aru Ranjan Singh , Abhishek Goswami , Kurt Debattista This is my paper · ORCID
classification cs.CV
keywords visible-to-infraredtranslationhighdynamicrangeimagingdensecaptioningvision-languagemodelsimage-to-imagepseudo-thermal-crossoverHDRTdataset
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

CapHDR2IR is a proposed framework for translating visible-light images into infrared images without specialized IR hardware. The paper's central claim is that two shortcomings explain why prior RGB-to-IR translation loses fidelity: standard dynamic range (SDR) visible input cannot represent extreme luminance, and the generator has no scene-level understanding, so objects with similar temperatures collapse into what the authors call pseudo-thermal-crossover artifacts. CapHDR2IR addresses both by taking high-dynamic-range (HDR) visible input and adding a dense caption branch, a pretrained vision-language model (ETDC) that emits per-region textual descriptions which are fused into the generator's encoder. On the HDRT dataset, the authors report the best PSNR, SSIM, MSE, and LPIPS in both the SDR-input and HDR-input comparison groups, and confirm the role of each component in an ablation study. If these results hold, visible-to-infrared synthesis becomes more practical for low-light and high-contrast scenes, because the input does the exposure work and the captions do the semantic work.

What carries the argument

The machinery is a two-branch generator. The caption branch tone-maps HDR input to SDR using the log-average luminance formula, then runs ETDC, a textual-context-aware dense captioner with a ResNet-101 backbone and region proposal network, to obtain region features and caption features. The image generation branch is an encoder-decoder: caption fusion modules use a spatial attention map computed from concatenated visual and caption features to align caption features with visual features at multiple scales, and up-sampling fusion modules concatenate center-cropped encoder features with transposed-convolution upsampled features. The training objective is perceptual loss from a pre-trained VGG network weighted at 10 plus a GAN loss weighted at 0.1.

What would settle it

Take an independently collected set of paired HDR-visible and infrared images, and compare the full CapHDR2IR against the same generator without the caption branch and against an SDR-input version, measuring whether distinct objects with similar temperatures remain separable in the output. If the caption branch offers no region-level improvement, or if SDR inputs match HDR inputs once the dataset is chosen to emphasize low-dynamic-range scenes, then the claim that HDR and captions drive the reported gains is falsified.

Watch

Extended reading notes

Core claim

The discovery the paper is trying to establish is that HDR input plus dense captions, not just a better generator architecture, is what moves visible-to-infrared translation forward. In the HDR group of Table 1, CapHDR2IR reports the best numbers on PSNR, SSIM, MSE, and LPIPS, beating RGB2IR, sRGB-TIR, CycleGAN, Pix2Pix, MUNIT, and classical methods on every metric. The authors argue that the HDR input preserves information in clipped shadows and highlights, while the caption branch retains objects that the naive generator would blur into dark, uniform regions that mimic thermal crossover. The ablation table supports the ordering: HDR input alone, the caption branch alone, and the caption fusion module each add measurable gains, and the full model with HDR pre-processing for the caption branch is the best configuration.

Load-bearing premise

The load-bearing premise is that the HDRT dataset is well-registered, representative, and unbiased in favor of HDR input or caption-friendly scenes; if that fails, the reported state-of-the-art may not hold on other visible-to-infrared benchmarks.

Editorial extensions

If this is right

  • Infrared generation pipelines should ingest HDR or multi-exposure visible images instead of single SDR frames when targeting dark and high-contrast scenes.
  • Dense captioning can act as a semantic prior that suppresses artifacts in domain transfer, a mechanism not limited to infrared imaging.
  • Each component of CapHDR2IR—HDR input, the caption branch, and the caption fusion module—contributes measurable improvement, and the full configuration is the only one that tops every metric on the HDRT benchmark.
  • The chosen loss weights (perceptual weight 10, GAN weight 0.1) place far more emphasis on feature-level fidelity than on adversarial realism, and the weight ablation identifies this as the best operating point.

Reading between the lines

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

  • A test the paper does not run: evaluating the same method on independently collected RGB-IR benchmarks would settle whether the HDR advantage generalizes beyond HDRT.
  • If the caption branch is doing the work, an ablation replacing dense captions with a segmentation or object-detection feature map would tell whether language is necessary or merely a source of high-level features.
  • A direct artifact metric—counting how often known distinct objects collapse to the same gray level in generated infrared output—would quantify pseudo-thermal crossover; the paper relies on qualitative examples.
  • Because the caption branch is pretrained on SDR imagery, the log-average tonemapping of HDR input is safety-critical, and any mismatch could silently degrade the semantic benefit.
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

4 major / 4 minor

Summary. The paper proposes CapHDR2IR, a two-branch generator for visible-to-infrared translation. One branch is an image generation encoder-decoder; the other is a dense captioning branch based on ETDC (Shao et al. 2023), whose features are fused into the encoder at multiple scales through caption fusion modules. The input is an HDR visible image, tone-mapped for the caption branch, and the training loss combines perceptual and GAN terms. Experiments on the HDRT dataset compare the method with traditional and deep baselines in both SDR-input and HDR-input settings, with ablations for HDR input, caption branch, caption fusion, and loss weights. The authors claim state-of-the-art performance on HDRT and argue that HDR input and dense captioning jointly address detail loss and pseudo-thermal-crossover artifacts.

Significance. If the reported results hold, the paper makes a useful contribution by identifying two concrete failure modes in visible-to-infrared translation and addressing them with HDR input and dense captions. The motivation is clearly argued, and the ablation study in Table 2 consistently shows gains from HDR input and from the caption branch, which supports the internal consistency of the design. The paper is also clearly written and the qualitative figures illustrate the intended effects. However, the central SOTA claim is currently supported only by a single self-collected benchmark, omits several directly relevant baselines, and relies on test-set-based hyperparameter selection; these issues prevent the claim from being taken at face value.

major comments (4)
  1. [Experiments, Comparison with Other Methods, Table 1] The comparison in Table 1 omits methods that the Related Work section explicitly identifies as visible-to-infrared translation approaches: PAS-GAN, InfraGAN, and IR-GAN. Because the text states that 'our model surpasses traditional methods and other deep learning approaches,' the absence of these three directly relevant baselines means the SOTA claim is not verifiable from the reported evidence. Please add these comparisons, or state concretely why they cannot be run (e.g., no public implementation), and condition the claim accordingly.
  2. [Loss Weight, Table 3] The loss weights alpha and beta in Eq. (1) are selected using the test set: Table 3 reports reconstruction metrics on the test set for different weight combinations, and the chosen values (alpha=10, beta=0.1) are then used for the main results. This makes the reported numbers in Table 1 the result of test-set selection. Please use a validation split for hyperparameter selection and report the corresponding test-set results, and include error bars or significance tests over multiple runs.
  3. [Experiments, Dataset] All experiments are conducted on HDRT, a dataset introduced by the same research group (Peng et al. 2024a), and no evaluation is performed on established RGB-IR benchmarks such as KAIST or FLIR. Since the central contribution depends on the claim that HDR inputs and caption fusion generalize across visible-to-infrared translation, the single-benchmark evaluation leaves the generalization claim unsupported. Please add at least one external dataset or explicitly restrict the conclusion to HDRT.
  4. [Ablation Study, Table 2] The improvements from the caption fusion module are small in Table 2 (e.g., PSNR 1.966 vs 1.976 for HDR2IRV3 vs CapHDR2IR), and no variance or significance information is reported anywhere in the ablation. Without repeated-run statistics, the marginal contribution of the final fusion module is difficult to distinguish from noise. Please provide standard deviations or significance tests for the ablation rows.
minor comments (4)
  1. [Methodology, Eq. (3)] Eq. (3) reuses alpha as a tone-mapping scaling factor, while Eq. (1) uses alpha as the perceptual-loss weight; this overloaded notation is confusing and should be resolved by renaming one of the two symbols.
  2. [Methodology, Eq. (7)] In Eq. (7), F^i_fused_cap is undefined; it should presumably be F^i_aligned_cap from Eq. (6).
  3. [Table 3] Table 3 is difficult to read because the alpha and beta values are not formatted as separate columns with clear headers (e.g., rows like 'e-1 e-1'); please reformat the table so that each hyperparameter value is explicit.
  4. [Experiments, Implementation Details] No code, trained weights, or detailed baseline training protocols are provided, which makes exact reproduction of Table 1 difficult; please consider releasing code and weights or providing more complete experimental details.

Circularity Check

0 steps flagged · score 0.0 of 10

No construction-level circularity; the SOTA claim is empirical, though self-referential evaluation choices weaken it.

full rationale

CapHDR2IR is an empirical systems paper; it does not derive a prediction from a first-principles equation. The claimed SOTA is supported by Table 1 on HDRT and by the ablations in Table 2. The main self-referential elements are (i) the HDRT dataset (Peng et al. 2024a), from overlapping authors, is the only benchmark; (ii) the dense caption backbone ETDC (Shao et al. 2023) is co-authored by two current authors and is adopted without comparison to other captioners; and (iii) the loss weights α=10 and β=0.1 (Eq. 1) were selected by comparing configurations in Table 3 on the same HDRT test set and then reported as final results. These are real limitations: they reduce external independence, risk test-set overfitting, and make the 'best weight' statement tautological over the tested grid. However, none of these makes a reported metric equal to an input by construction: the α/β choice does not force CapHDR2IR to beat other methods, and the ablations separate the contributions of HDR input and caption fusion. The omission of PAS-GAN, InfraGAN, and IR-GAN from Table 1 despite their mention in Related Work is a comparison-completeness risk, not circularity. Under the rule that circularity requires a specific equation-level reduction or renamed fit, no such step is present.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on one self-authored dataset and a frozen external caption backbone; the only fitted numbers are the loss weights and an unspecified tone-mapping scale. No new physical entities are introduced.

free parameters (3)
  • perceptual loss weight alpha = 10
    Chosen via ablation on the test set (Table 3); directly affects the training objective and all reported metrics.
  • GAN loss weight beta = 0.1
    Chosen via ablation on the test set (Table 3); directly affects the training objective and all reported metrics.
  • tone-mapping scaling factor in Eq. (3) = not specified
    Controls brightness of the SDR conversion for the caption branch; no value is given, making the preprocessing hard to reproduce.
assumptions (4)
  • domain assumption HDRT dataset pairs are well-registered and representative of real visible/IR conditions.
    All evaluation and ablations rely on this dataset (section 'Dataset'); if the alignment or sensor response is biased, the conclusions may not generalize.
  • domain assumption The pretrained ETDC dense caption model provides semantically correct features when applied to tone-mapped HDR images.
    The caption branch is used as a frozen feature extractor after log-average tone mapping (Eqs. 2-4); no fine-tuning or caption accuracy evaluation is reported.
  • standard math VGG-based perceptual loss is a valid fidelity measure for infrared images.
    Perceptual loss is used as the main training signal; this is standard in image generation but not physically validated for IR.
  • domain assumption The combination of perceptual and GAN losses with fixed weights yields a stable training objective.
    No training instability is reported, but the weights were tuned on the test set.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CapHDR2IR: Caption-Driven Transfer from Visible Light to Infrared Domain." pith.science (2026). https://pith.science/paper/AS3MKX6H

@misc{pith2026241116327,
  author       = {Pith},
  title        = {Pith review of: CapHDR2IR: Caption-Driven Transfer from Visible Light to Infrared Domain},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AS3MKX6H}},
  note         = {Machine review of arXiv:2411.16327}
}
read the original abstract

Infrared (IR) imaging offers advantages in several fields due to its unique ability of capturing content in extreme light conditions. However, the demanding hardware requirements of high-resolution IR sensors limit its widespread application. As an alternative, visible light can be used to synthesize IR images but this causes a loss of fidelity in image details and introduces inconsistencies due to lack of contextual awareness of the scene. This stems from a combination of using visible light with a standard dynamic range, especially under extreme lighting, and a lack of contextual awareness can result in pseudo-thermal-crossover artifacts. This occurs when multiple objects with similar temperatures appear indistinguishable in the training data, further exacerbating the loss of fidelity. To solve this challenge, this paper proposes CapHDR2IR, a novel framework incorporating vision-language models using high dynamic range (HDR) images as inputs to generate IR images. HDR images capture a wider range of luminance variations, ensuring reliable IR image generation in different light conditions. Additionally, a dense caption branch integrates semantic understanding, resulting in more meaningful and discernible IR outputs. Extensive experiments on the HDRT dataset show that the proposed CapHDR2IR achieves state-of-the-art performance compared with existing general domain transfer methods and those tailored for visible-to-infrared image translation.

Figures

Figures reproduced from arXiv: 2411.16327 by the authors.

Figure 1
Figure 1. Motivation of the proposed CapHDR2IR model. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The overall structure of the proposed CapHDR2IR. The process comprises two branches: a caption branch (upper [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 2
Figure 2. The caption branch aims to distill the content of the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figures from the paper (3 more)
Figure 3
Figure 3. Figure 3: Visualization results on the HDRT dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]
Figure 4
Figure 4. Figure 4: Comparison of generation performance using SDR vs. HDR inputs. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Comparison of generation performance with or without the caption branch. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 38 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    G.; Duddu, H.; Shirtliffe, S.; Vail, S.; Bett, K.; Pozniak, C.; and Stavness, I

    Aslahishahri, M.; Stanley, K. G.; Duddu, H.; Shirtliffe, S.; Vail, S.; Bett, K.; Pozniak, C.; and Stavness, I. 2021. From RGB to NIR: Predicting of near infrared reflectance from visible spectrum aerial images of crops. In 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 1312--1322

    work page 2021

  4. [4]

    Boroujeni, S. P. H.; and Razi, A. 2024. IC-GAN: An Improved Conditional Generative Adversarial Network for RGB-to-IR image translation with applications to forest fire monitoring. Expert Systems with Applications, 238: 121962

    work page 2024

  5. [5]

    Castleman, K. R. 1996. Digital image processing. USA: Prentice Hall Press. ISBN 0132114674

    work page 1996

  6. [6]

    C.; Saqui, D.; Ataky, S.; Jorge, L

    de Lima, D. C.; Saqui, D.; Ataky, S.; Jorge, L. A. d. C.; Ferreira, E. J.; and Saito, J. H. 2019. Estimating Agriculture NIR Images from Aerial RGB Data. In Rodrigues, J. M. F.; Cardoso, P. J. S.; Monteiro, J.; Lam, R.; Krzhizhanovskaya, V. V.; Lees, M. H.; Dongarra, J. J.; and Sloot, P. M., eds., Computational Science -- ICCS 2019, 562--574. Cham: Spring...

    work page 2019

  7. [7]

    Gonzalez, R.; and Woods, R. 2008. Digital Image Processing. Prentice Hall. ISBN 9780131687288

    work page 2008

  8. [8]

    Haefner, B.; Green, S.; Oursland, A.; Andersen, D.; Goesele, M.; Cremers, D.; Newcombe, R.; and Whelan, T. 2021. Recovering Real-World Reflectance Properties and Shading From HDR Imagery. In 2021 International Conference on 3D Vision (3DV), 1075--1084

    work page 2021

Show all 41 references
  1. [9]

    Hou, S.; Lindsay, C.; Agu, E.; Pedersen, P.; Tulu, B.; and Strong, D. 2021. HDR-Like Image Generation to Mitigate Adverse Wound Illumination Using Deep Bi-directional Retinex and Exposure Fusion. In Papie \. z , B. W.; Yaqub, M.; Jiao, J.; Namburete, A. I. L.; and Noble, J. A....

    2021

  2. [10]

    Huang, F.; Huang, W.; and Wu, X. 2024. Enhancing Infrared Optical Flow Network Computation through RGB-IR Cross-Modal Image Generation. Sensors, 24(5)

    2024

  3. [11]

    Huang, X.; Liu, M.-Y.; Belongie, S.; and Kautz, J. 2018. Multimodal Unsupervised Image-to-Image Translation. In Ferrari, V.; Hebert, M.; Sminchisescu, C.; and Weiss, Y., eds., Computer Vision -- ECCV 2018, 179--196. Cham: Springer International Publishing. ISBN 978-3-030-01219-9

    2018

  4. [12]

    Isola, P.; Zhu, J.-Y.; Zhou, T.; and Efros, A. A. 2017. Image-to-Image Translation with Conditional Adversarial Networks. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 5967--5976

    2017

  5. [13]

    Johnson, J.; Karpathy, A.; and Fei-Fei, L. 2016. DenseCap: Fully Convolutional Localization Networks for Dense Captioning. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 4565--4574

    2016

  6. [14]

    Kou, R.; Wang, C.; Peng, Z.; Zhao, Z.; Chen, Y.; Han, J.; Huang, F.; Yu, Y.; and Fu, Q. 2023. Infrared small target segmentation networks: A survey. Pattern Recognition, 143: 109788

    2023

  7. [15]

    Lee, D.-G.; Jeon, M.-H.; Cho, Y.; and Kim, A. 2023. Edge-guided Multi-domain RGB-to-TIR image Translation for Training Vision Tasks with Challenging Labels. In 2023 IEEE International Conference on Robotics and Automation (ICRA), 8291--8298

    2023

  8. [16]

    Li, H.; Xu, T.; Wu, X.-J.; Lu, J.; and Kittler, J. 2023. LRRNet: A Novel Representation Learning Guided Fusion Network for Infrared and Visible Images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(9): 11040--11052

    2023

  9. [17]

    Li, X.; Jiang, S.; and Han, J. 2019. Learning Object Context for Dense Captioning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, 8650--8657

    2019

  10. [18]

    Li, X.; Wen, C.; Hu, Y.; Yuan, Z.; and Zhu, X. X. 2024. Vision-Language Models in Remote Sensing: Current progress and future trends. IEEE Geoscience and Remote Sensing Magazine, 12(2): 32--66

    2024

  11. [19]

    Ma, D.; Xian, Y.; Li, B.; Li, S.; and Zhang, D. 2024. Visible-to-infrared image translation based on an improved CGAN. VISUAL COMPUTER, 40(2): 1289--1298

    2024

  12. [20]

    Peng, J.; Bashford-Rogers, T.; Banterle, F.; Zhao, H.; and Debattista, K. 2024 a . HDRT: Infrared Capture for HDR Imaging. arXiv preprint, 2406.05475

    2024 arXiv

  13. [21]

    Peng, J.; Zhao, H.; and Hu, Z. 2023. Dynamic Fusion Network for RGBT Tracking. IEEE Transactions on Intelligent Transportation Systems, 24(4): 3822--3832

    2023

  14. [22]

    Peng, J.; Zhao, H.; Hu, Z.; Zhuang, Y.; and Wang, B. 2023 a . Siamese infrared and visible light fusion network for RGB-T tracking. International Journal of Machine Learning and Cybernetics, 14(9): 3281--3293

    2023

  15. [23]

    Peng, J.; Zhao, H.; Zhao, K.; Wang, Z.; and Yao, L. 2023 b . CourtNet: Dynamically balance the precision and recall rates in infrared small target detection. Expert Systems with Applications, 233: 120996

    2023

  16. [24]

    Peng, J.; Zhao, H.; Zhao, K.; Wang, Z.; and Yao, L. 2024 b . Dynamic background reconstruction via masked autoencoders for infrared small target detection. Engineering Applications of Artificial Intelligence, 135: 108762

    2024

  17. [25]

    Reinhard, E.; Adhikhmin, M.; Gooch, B.; and Shirley, P. 2001. Color transfer between images. IEEE Computer Graphics and Applications, 21(5): 34--41

    2001

  18. [26]

    Seetzen, H.; Heidrich, W.; Stuerzlinger, W.; Ward, G.; Whitehead, L.; Trentacoste, M.; Ghosh, A.; and Vorozcovs, A. 2004. High dynamic range display systems. ACM Trans. Graph., 23(3): 760–768

    2004

  19. [27]

    R.; and Jalal, A

    Shahzad, A. R.; and Jalal, A. 2021. A Smart Surveillance System for Pedestrian Tracking and Counting using Template Matching. In 2021 International Conference on Robotics and Automation in Industry (ICRAI), 1--6

    2021

  20. [28]

    Shao, Z.; Han, J.; Debattista, K.; and Pang, Y. 2023. Textual Context-Aware Dense Captioning With Diverse Words. IEEE Transactions on Multimedia, 25: 8753--8766

    2023

  21. [29]

    Shao, Z.; Han, J.; Marnerides, D.; and Debattista, K. 2022. Region-Object Relation-Aware Dense Captioning via Transformer. IEEE Transactions on Neural Networks and Learning Systems, 1--12

    2022

  22. [30]

    Shopovska, I.; Stojkovic, A.; Aelterman, J.; Van Hamme, D.; and Philips, W. 2023. High-Dynamic-Range Tone Mapping in Intelligent Automotive Systems. Sensors, 23(12)

    2023

  23. [31]

    Shukla, A.; Upadhyay, A.; Sharma, M.; Chinnusamy, V.; and Kumar, S. 2022. High-Resolution NIR Prediction from RGB Images: Application to Plant Phenotyping. In 2022 IEEE International Conference on Image Processing (ICIP), 4058--4062

    2022

  24. [32]

    R.; Bashford-Rogers, T.; Marnerides, D.; Debattista, K.; and Hazra, S

    Singh, A. R.; Bashford-Rogers, T.; Marnerides, D.; Debattista, K.; and Hazra, S. 2023. HDR image-based deep learning approach for automatic detection of split defects on sheet metal stamping parts. The International Journal of Advanced Manufacturing Technology, 125(5): 2393--2408

    2023

  25. [33]

    Tang, L.; Xiang, X.; Zhang, H.; Gong, M.; and Ma, J. 2023. DIVFusion: Darkness-free infrared and visible image fusion. Information Fusion, 91: 477--493

    2023

  26. [34]

    Wang, S.; Sun, G.; Dong, L.; and Zheng, B. 2024. PAS-GAN: A GAN based on the Pyramid Across-Scale module for visible-infrared image transformation. Infrared Physics & Technology, 139: 105314

    2024

  27. [35]

    Wu, H.; Hao, X.; Wu, J.; Xiao, H.; He, C.; and Yin, S. 2023. Deep learning-based image super-resolution restoration for mobile infrared imaging system. Infrared Physics & Technology, 132: 104762

    2023

  28. [36]

    Yang, L.; Tang, K.; Yang, J.; and Li, L.-J. 2017. Dense Captioning with Joint Inference and Visual Context. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 1978--1987

    2017

  29. [37]

    Yin, G.; Sheng, L.; Liu, B.; Yu, N.; Wang, X.; and Shao, J. 2019. Context and Attribute Grounded Dense Captioning. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 6234--6243

    2019

  30. [38]

    Zhang, J.; Huang, J.; Jin, S.; and Lu, S. 2024. Vision-Language Models for Vision Tasks: A Survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(8): 5625--5644

    2024

  31. [39]

    Zhang, L.; Xiao, M.; Wang, Y.; Peng, S.; Chen, Y.; Zhang, D.; Zhang, D.; Guo, Y.; Wang, X.; Luo, H.; Zhou, Q.; and Xu, Y. 2021. Fast Screening and Primary Diagnosis of COVID-19 by ATR–FT-IR. Analytical Chemistry, 93(4): 2191--2199. PMID: 33427452

    2021

  32. [40]

    Zhu, J.-Y.; Park, T.; Isola, P.; and Efros, A. A. 2017. Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks. In 2017 IEEE International Conference on Computer Vision (ICCV), 2242--2251

    2017

  33. [41]

    A.; and Ozer, S

    Özkanoğlu, M. A.; and Ozer, S. 2022. InfraGAN: A GAN architecture to transfer visible images to infrared domain. Pattern Recognition Letters, 155: 69--76

    2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.