pith. sign in

arxiv: 2605.20766 · v1 · pith:ZDAUQ3XTnew · submitted 2026-05-20 · 💻 cs.CV

Diffuse to Detect: Bi-Level Sample Rebalancing with Pseudo-Label Diffusion for Point-Supervised Infrared Small-Target Detection

Zhu Liu , Yuanhang Yao , Ping Qian , Zihang Chen , Risheng Liu This is my paper

Pith reviewed 2026-05-21 05:53 UTC · model grok-4.3

classification 💻 cs.CV
keywords infrared small target detectionpoint supervisionpseudo label diffusionsample rebalancingbi-level optimizationphysics induced annotationmeta classifierdifferentiable diffusion module
0
0 comments X

The pith

Physics-based diffusion converts single-point labels to reliable pseudo-masks for infrared small-target detection.

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

The paper seeks to overcome unstable pseudo-label evolution and sample imbalance in point-supervised infrared small target detection. It establishes that the consistency between thermal radiation patterns and heat diffusion can be used to expand point labels into trustworthy pseudo-masks. A bi-level framework then allows joint optimization of the detector, sample weights, and diffusion parameters for adaptive training. A sympathetic reader would care because this promises much faster annotation and effective performance even with limited data in difficult imaging conditions.

Core claim

Leveraging the intrinsic consistency between thermal radiation patterns and heat diffusion, a physics-induced annotation strategy expands single-point labels into reliable pseudo-masks. A bi-level dual-update framework jointly optimizes detector weights, sample weights predicted by a meta-classifier, and diffusion parameters in a differentiable module that uses detection feedback to refine the pseudo-labels.

What carries the argument

bi-level dual-update framework incorporating a meta-classifier for sample-wise loss weights and a differentiable diffusion module that refines pseudo-labels using detection feedback

Load-bearing premise

An intrinsic consistency between thermal radiation patterns and heat diffusion is strong enough to reliably convert single-point labels into accurate pseudo-masks within cluttered low-contrast infrared imagery.

What would settle it

Comparing the pseudo-masks produced by the diffusion process against ground-truth target masks on a held-out set of fully annotated infrared images; if the average overlap or precision is no better than a simple dilation of the point labels, the core consistency assumption would be falsified.

Figures

Figures reproduced from arXiv: 2605.20766 by Ping Qian, Risheng Liu, Yuanhang Yao, Zhu Liu, Zihang Chen.

Figure 1
Figure 1. Figure 1: Motivation and efficiency overview. (a) Illustration of challenges including sample imbalance and complexity. (b) Quantitative results showing clear improvements using only 30% of the data. (c) Qualitative results validating our robustness in complex scenes. pseudo-labels through dedicated algorithms (Li et al., 2023a; Kou et al., 2024), such as MCLC (Li et al., 2023a), which constructs labels from repeate… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the proposed framework. (a) Bi-level dual-update framework performs joint sample rebalancing and label refinement. (b) Dynamic sample rebalancing is designed to weight training data. (c) Physics-induced diffusion annotation generates reliable pseudo-masks from single-point supervision in a learnable manner. However, these methods require careful model design and fine-tuning of hyperparameters, … view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative comparison on the SIRST3 dataset. Columns from left to right: 3D surfaces of inputs, input images with labels and local magnification, Ours, LESPS, MCLC, and PAL predictions (the blue part on ALCLNet, the green part on DNANet) [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparison on three representative scenes. These results are obtained by ALCLNet. tures (Rows 1 and 2), competing methods struggle signifi￾cantly. A multi-target case is shown in the third row, where our method separates targets cleanly and avoids merging them into surrounding structures, while preserving target boundaries more effectively overall. A low-contrast multi￾target scenario is shown … view at source ↗
Figure 6
Figure 6. Figure 6: Data-efficiency on SIRST3 dataset across diverse net￾works, comparing Ours, LESPS, PAL, and MCLC under single￾point supervision. Others are trained with full training datasets. Epoch (a) Selection Strategy Comparison IoU IoU Seed (b) Stability Validation [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Convergence of sample-selection strategy and stability validation on SIRST3 under diverse seeds. compares three data selection strategies under the same data budget. Our learned weights clearly outperform random and prior-based (hand-crafted difficulty) selection, showing more reliable identification of informative samples [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative comparison of small target detection on the SIRST3 dataset: Ours vs. COM, MCLC, SAM, and SAM2. speedup over MCLC) is crucial: it makes the diffusion mod￾ule lightweight enough to be embedded within the bi-level optimization loop. We compared our proposed physics￾diffusion annotation with foundation models SAM/SAM2 in [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
read the original abstract

Point supervision has become a scalable solution to address dense annotation for infrared small target detection, but its performance is limited by two coupled bottlenecks: unstable pseudo-label evolution in cluttered, low-contrast infrared imagery and severe sample-distribution imbalance. In this paper, we present a more adaptive and stable framework to address these issues. Leveraging the intrinsic consistency between thermal radiation patterns and heat diffusion, we propose a physics-induced annotation strategy that expands single-point labels into reliable pseudo-masks. To further enhance supervision and alleviate sample imbalance, we develop a bi-level dual-update framework that jointly optimizes detector weights, sample weights, and diffusion parameters. A meta-classifier dynamically predicts sample-wise loss weights, while a differentiable diffusion module refines pseudo-labels with detection feedback, enabling adaptive interaction between training and hyperparameter optimization. Extensive experiments across multiple datasets demonstrate five-fold annotation acceleration, superior detection accuracy, and comparable performance with 30% of the training data, validating the efficiency and practicality of our approach. Our code is available at https://github.com/yuanhang-yao/diffuse-to-detect.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes 'Diffuse to Detect', a bi-level framework for point-supervised infrared small-target detection. It introduces a physics-induced annotation strategy that expands single-point labels into pseudo-masks by modeling heat diffusion, justified by an assumed intrinsic consistency between thermal radiation patterns and heat diffusion in cluttered low-contrast scenes. A meta-classifier predicts sample-wise loss weights to address imbalance, while a differentiable diffusion module refines the pseudo-labels using detection feedback, enabling joint optimization of detector weights, sample weights, and diffusion parameters. Experiments across datasets are reported to demonstrate five-fold annotation acceleration, superior detection accuracy, and comparable performance using only 30% of the training data.

Significance. If the generated pseudo-masks prove reliable, the method could meaningfully reduce annotation effort for infrared small-target detection, a domain where dense labeling is costly. The combination of physics-inspired diffusion with bi-level optimization for adaptive pseudo-label refinement and sample rebalancing offers a potentially useful direction for weakly-supervised CV tasks. Open-sourced code supports reproducibility and allows independent verification of the claimed efficiency gains.

major comments (2)
  1. [Abstract / Method description] The central claim depends on the diffusion module producing reliable pseudo-masks that approximate true target shapes rather than introducing systematic errors. The abstract and method description treat the 'intrinsic consistency between thermal radiation patterns and heat diffusion' as given, yet no direct quantitative validation (e.g., mask IoU or overlap metrics against held-out full annotations in cluttered scenes) is reported to confirm this in the target domain; without such evidence the performance gains cannot be attributed to the physics-induced strategy.
  2. [Bi-level optimization framework] The bi-level dual-update framework jointly optimizes detector weights, meta-classifier sample weights, and diffusion parameters. This creates a risk that reported improvements partly reflect fitting to the diffusion hyperparameters rather than independent generalization; the manuscript should include an ablation isolating the contribution of the differentiable diffusion module versus post-hoc tuning.
minor comments (2)
  1. [Figures] Add visual comparisons in the figures showing generated pseudo-masks overlaid on original IR images versus ground-truth masks to illustrate behavior in low-contrast cluttered regions.
  2. [Method] Clarify the exact loss formulation and how detection feedback is back-propagated through the diffusion module to ensure the interaction loop is fully differentiable.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We address each major comment below and indicate the revisions we will make to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Abstract / Method description] The central claim depends on the diffusion module producing reliable pseudo-masks that approximate true target shapes rather than introducing systematic errors. The abstract and method description treat the 'intrinsic consistency between thermal radiation patterns and heat diffusion' as given, yet no direct quantitative validation (e.g., mask IoU or overlap metrics against held-out full annotations in cluttered scenes) is reported to confirm this in the target domain; without such evidence the performance gains cannot be attributed to the physics-induced strategy.

    Authors: We agree that direct quantitative validation of the pseudo-masks would provide stronger support for the physics-induced strategy. Although downstream detection metrics and comparisons offer indirect validation, we will add an analysis computing mask IoU and overlap metrics between the generated pseudo-masks and available ground-truth annotations on evaluation subsets where full labels exist. This will be included in the revised manuscript to better attribute performance gains to the diffusion approach. revision: yes

  2. Referee: [Bi-level optimization framework] The bi-level dual-update framework jointly optimizes detector weights, meta-classifier sample weights, and diffusion parameters. This creates a risk that reported improvements partly reflect fitting to the diffusion hyperparameters rather than independent generalization; the manuscript should include an ablation isolating the contribution of the differentiable diffusion module versus post-hoc tuning.

    Authors: We recognize the value of isolating the contribution of the differentiable diffusion module. The current ablations focus on the overall bi-level framework, but we will add a new experiment in the revision that compares the joint bi-level optimization against a post-hoc tuned diffusion variant with fixed parameters. This will clarify whether the adaptive feedback provides benefits beyond hyperparameter fitting. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on external physical assumption and standard bi-level optimization

full rationale

The paper's core claim rests on an assumed 'intrinsic consistency between thermal radiation patterns and heat diffusion' to expand point labels into pseudo-masks, presented as a physics-induced prior rather than any self-referential definition or fitted input renamed as prediction. The bi-level dual-update framework jointly optimizes detector weights, sample weights, and diffusion parameters via a meta-classifier and differentiable module, but this constitutes a conventional adaptive training loop with detection feedback; no equations or steps in the provided description reduce a claimed result (e.g., reliable pseudo-masks or performance gains) to its own inputs by construction. No self-citations, uniqueness theorems, or ansatzes imported from prior author work are referenced as load-bearing. Experiments on multiple datasets provide external validation, keeping the approach self-contained against benchmarks without circular reduction.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The approach rests on one domain assumption about thermal diffusion consistency and introduces diffusion parameters that are optimized inside the training loop; no explicit free parameters or invented physical entities are named in the abstract.

free parameters (1)
  • diffusion parameters
    Jointly optimized with detector and sample weights in the bi-level framework; exact values not stated in abstract.
axioms (1)
  • domain assumption Intrinsic consistency between thermal radiation patterns and heat diffusion allows reliable expansion of point labels into pseudo-masks
    Invoked to justify the physics-induced annotation strategy in the abstract.

pith-pipeline@v0.9.0 · 5736 in / 1378 out tokens · 30334 ms · 2026-05-21T05:53:34.933937+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

  • IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    Leveraging the intrinsic consistency between thermal radiation patterns and heat diffusion, we propose a physics-induced annotation strategy that expands single-point labels into reliable pseudo-masks... differentiable diffusion module refines pseudo-labels

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

80 extracted references · 80 canonical work pages

  1. [1]

    Yu, Yi and Ren, Botao and Zhang, Peiyuan and Liu, Mingxin and Luo, Junwei and Zhang, Shaofeng and Da, Feipeng and Yan, Junchi and Yang, Xue , booktitle=

    work page

  2. [2]

    2025 , publisher=

    Yu, Yi and Yang, Xue and Li, Yansheng and Han, Zhenjun and Da, Feipeng and Yan, Junchi , journal=. 2025 , publisher=

    work page 2025

  3. [3]

    Luo, Junwei and Yang, Xue and Yu, Yi and Li, Qingyun and Yan, Junchi and Li, Yansheng , booktitle=

    work page

  4. [4]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Investigating bi-level optimization for learning and vision from a unified perspective: A survey and beyond , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2022 , publisher=

    work page 2022

  5. [5]

    Advances in Neural Information Processing Systems , volume=

    Bilevel Optimization for Adversarial Learning Problems: Sharpness, Generation, and Beyond , author=. Advances in Neural Information Processing Systems , volume=

    work page

  6. [6]

    Wang and D

    Y. Wang and D. Jin and J. Chen and X. Bai. Revelation of hidden 2D atmospheric turbulence strength fields from turbulence effects in infrared imaging. Nature Computational Science. 2023

    work page 2023

  7. [7]

    Bao and X

    F. Bao and X. Wang and S. H. Sureshbabu and G. Sreekumar and L. Yang and V. Aggarwal and V. N. Boddeti and Z. Jacob. Heat-assisted detection and ranging. Nature. 2023

    work page 2023

  8. [8]

    Kou and C

    R. Kou and C. Wang and Q. Fu and Z. Li and Y. Luo and B. Li and W. Li and Z. Peng. MCGC : A multiscale chain growth clustering algorithm for generating infrared small target mask under single-point supervision. IEEE Transactions on Geoscience and Remote Sensing. 2024

    work page 2024

  9. [9]

    Advances in Neural Information Processing Systems , volume=

    Beyond value functions: Single-loop bilevel optimization under flatness conditions , author=. Advances in Neural Information Processing Systems , volume=

    work page

  10. [10]

    Liu and Z

    R. Liu and Z. Liu and J. Liu and X. Fan and Z. Luo. A task-guided, implicitly-searched and meta-initialized deep model for image fusion. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2024

    work page 2024

  11. [11]

    Li and C

    B. Li and C. Xiao and L. Wang and Y. Wang and Z. Lin and M. Li and W. An and Y. Guo. Dense nested attention network for infrared small target detection. IEEE Transactions on Image Processing. 2023

    work page 2023

  12. [12]

    Driggers and E

    R. Driggers and E. Pollak and R. Grimming and E. Velazquez and R. Short and G. Holst and O. Furxhi. Detection of small targets in the infrared: an infrared search and track tutorial. Applied Optics. 2021

    work page 2021

  13. [13]

    Xu and S

    H. Xu and S. Zhong and T. Zhang and X. Zou. Multiscale multilevel residual feature fusion for real-time infrared small target detection. IEEE Transactions on Geoscience and Remote Sensing. 2023

    work page 2023

  14. [14]

    Zhao and L

    E. Zhao and L. Dong and C. Li and Y. Ji. Infrared maritime target detection based on temporal weight and total variation regularization under strong wave interferences. IEEE Transactions on Geoscience and Remote Sensing. 2024

    work page 2024

  15. [15]

    Zhao and C

    J. Zhao and C. Yu and Z. Shi and Y. Liu and Y. Zhang. Gradient-guided learning network for infrared small target detection. IEEE Geoscience and Remote Sensing Letters. 2023

    work page 2023

  16. [16]

    Chen and S

    Z. Chen and S. Luo and T. Xie and J. Liu and G. Wang and G. Lei. A novel infrared small target detection method based on BEMD and local inverse entropy. Infrared Physics & Technology. 2014

    work page 2014

  17. [17]

    Dai and Y

    Y. Dai and Y. Wu and Y. Song and J. Guo. Non-negative infrared patch-image model: Robust target-background separation via partial sum minimization of singular values. Infrared Physics & Technology. 2017

    work page 2017

  18. [18]

    Deng and X

    H. Deng and X. Sun and M. Liu and C. Ye and X. Zhou. Small infrared target detection based on weighted local difference measure. IEEE Transactions on Geoscience and Remote Sensing. 2016

    work page 2016

  19. [19]

    Gao and D

    C. Gao and D. Meng and Y. Yang and Y. Wang and X. Zhou and A. G. Hauptmann. Infrared patch-image model for small target detection in a single image. IEEE Transactions on Image Processing. 2013

    work page 2013

  20. [20]

    He and M

    Y. He and M. Li and J. Zhang and Q. An. Small infrared target detection based on low-rank and sparse representation. Infrared Physics & Technology. 2015

    work page 2015

  21. [21]

    Qi and T

    B. Qi and T. Wu and H. He. Robust detection of small infrared objects in maritime scenarios using local minimum patterns and spatio-temporal context. Optical Engineering. 2012

    work page 2012

  22. [22]

    Qin and L

    Y. Qin and L. Bruzzone and C. Gao and B. Li. Infrared small target detection based on facet kernel and random walker. IEEE Transactions on Geoscience and Remote Sensing. 2019

    work page 2019

  23. [23]

    Wei and X

    Y. Wei and X. You and H. Li. Multiscale patch-based contrast measure for small infrared target detection. Pattern Recognition. 2016

    work page 2016

  24. [24]

    Zhang and L

    L. Zhang and L. Peng and T. Zhang and S. Cao and Z. Peng. Infrared small target detection via non-convex rank approximation minimization joint l 2, 1 norm. Remote Sensing. 2018

    work page 2018

  25. [25]

    Chen and L

    S. Chen and L. Ji and S. Peng and S. Zhu and M. Ye and Y. Sang. Language-Driven Motion Prior Knowledge Learning for Moving Infrared Small Target Detection. IEEE Transactions on Geoscience and Remote Sensing. 2025

    work page 2025

  26. [26]

    Chen and W

    G. Chen and W. Wang and S. Tan. Irstformer: A hierarchical vision transformer for infrared small target detection. Remote Sensing. 2022

    work page 2022

  27. [27]

    Qiu and Y

    Z. Qiu and Y. Ma and F. Fan and J. Huang and M. Wu and X. Mei. A pixel-level local contrast measure for infrared small target detection. Defence Technology. 2022

    work page 2022

  28. [28]

    Arce and M

    G. Arce and M. McLoughlin. Theoretical analysis of the max/median filter. IEEE Transactions on Acoustics, Speech, and Signal Processing. 1987

    work page 1987

  29. [29]

    Bai and Y

    X. Bai and Y. Bi. Derivative entropy-based contrast measure for infrared small-target detection. IEEE Transactions on Geoscience and Remote Sensing. 2018

    work page 2018

  30. [30]

    Bai and F

    X. Bai and F. Zhou. Analysis of new top-hat transformation and the application for infrared dim small target detection. Pattern Recognition. 2010

    work page 2010

  31. [31]

    C. L. P. Chen and H. Li and Y. Wei and T. Xia and Y. Y. Tang. A local contrast method for small infrared target detection. IEEE Transactions on Geoscience and Remote Sensing. 2014

    work page 2014

  32. [32]

    Yang and X

    B. Yang and X. Zhang and J. Zhang and J. Luo and M. Zhou and Y. Pi. EFLNet: Enhancing feature learning network for infrared small target detection. IEEE Transactions on Geoscience and Remote Sensing. 2024

    work page 2024

  33. [33]

    Zhao and W

    M. Zhao and W. Li and L. Li and J. Hu and P. Ma and R. Tao. Single-frame infrared small-target detection: A survey. IEEE Geoscience and Remote Sensing Magazine. 2022

    work page 2022

  34. [34]

    Lin and P

    S. Lin and P. Wang and M. Luo. Control traffic balancing in software defined networks. Computer Networks. 2016

    work page 2016

  35. [35]

    Dai and Y

    Y. Dai and Y. Wu and F. Zhou and K. Barnard. Attentional local contrast networks for infrared small target detection. IEEE Transactions on Geoscience and Remote Sensing. 2021

    work page 2021

  36. [36]

    Wu and D

    X. Wu and D. Hong and J. Chanussot. UIU-Net : U-Net in U-Net for infrared small object detection. IEEE Transactions on Image Processing. 2023

    work page 2023

  37. [37]

    Yu and Y

    C. Yu and Y. Liu and S. Wu and X. Xia and Z. Hu and D. Lan and X. Liu. Pay attention to local contrast learning networks for infrared small target detection. IEEE Geoscience and Remote Sensing Letters. 2022

    work page 2022

  38. [38]

    Zhu and S

    H. Zhu and S. Liu and L. Deng and Y. Li and F. Xiao. Infrared small target detection via low-rank tensor completion with top-hat regularization. IEEE Transactions on Geoscience and Remote Sensing. 2020

    work page 2020

  39. [39]

    Hou and Z

    Q. Hou and Z. Wang and F. Tan and Y. Zhao and H. Zheng and W. Zhang. RISTDnet: Robust infrared small target detection network. IEEE Geoscience and Remote Sensing Letters. 2021

    work page 2021

  40. [40]

    Li and Y

    R. Li and Y. Shen. YOLOSR-IST: A deep learning method for small target detection in infrared remote sensing images based on super-resolution and YOLO. Signal Processing. 2023

    work page 2023

  41. [41]

    Lu and Y

    Y. Lu and Y. Lin and H. Wu and X. Xian and Y. Shi and L. Lin. SIRST-5K : Exploring massive negatives synthesis with self-supervised learning for robust infrared small target detection. IEEE Transactions on Geoscience and Remote Sensing. 2024

    work page 2024

  42. [42]

    Lu and T

    X. Lu and T. Yue and J. Cai and Y. Chen and C. Lv and S. Chu. MSCA-Net : Multi-scale context aggregation network for infrared small target detection. Optics & Laser Technology. 2025

    work page 2025

  43. [43]

    Zhao and Z

    J. Zhao and Z. Shi and C. Yu and Y. Liu and X. Ying and Y. Dai. Multi-Scale Direction-Aware Network for Infrared Small Target Detection. IEEE Transactions on Geoscience and Remote Sensing. 2025

    work page 2025

  44. [44]

    Li and M

    Q. Li and M. Zhang and Z. Yang and Y. Yuan and Q. Wang. Edge-guided perceptual network for infrared small target detection. IEEE Transactions on Geoscience and Remote Sensing. 2024

    work page 2024

  45. [45]

    Spatial–Frequency Domain Transformation for Infrared Small Target Detection , year=

    Liu, Yuhao and Tu, Bing and Liu, Bo and He, Yan and Li, Jun and Plaza, Antonio , journal=. Spatial–Frequency Domain Transformation for Infrared Small Target Detection , year=

    work page

  46. [46]

    Yu and Y

    C. Yu and Y. Liu and S. Wu and Z. Hu and X. Xia and D. Lan and X. Liu. Infrared small target detection based on multiscale local contrast learning networks. Infrared Physics & Technology. 2022

    work page 2022

  47. [47]

    Loshchilov and F

    I. Loshchilov and F. Hutter. Decoupled weight decay regularization. 2017

    work page 2017

  48. [48]

    Liu and M

    X. Liu and M. Zhu and L. Lu and H. Sun and J. Wang. Multi-resolution partial differential equations preserved learning framework for spatiotemporal dynamics. Communications Physics. 2024

    work page 2024

  49. [49]

    Jiang and P

    C. Jiang and P. Kilcullen and Y. Lai and T. Ozaki and J. Liang. Single-pixel infrared imaging thermometry maps human inner canthi temperature. Nature Communications. 2025

    work page 2025

  50. [50]

    Perona and J

    P. Perona and J. Malik. Scale-space and Edge Detection Using Anisotropic Diffusion. IEEE Transactions on Pattern Analysis and Machine Intelligence. 1990. doi:10.1109/34.56205

    work page doi:10.1109/34.56205 1990

  51. [51]

    Liu and G

    R. Liu and G. Zhong and J. Cao and Z. Lin and S. Shan and Z. Luo. Learning to diffuse: A new perspective to design PDEs for visual analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2016

    work page 2016

  52. [52]

    Kirillov and E

    A. Kirillov and E. Mintun and N. Ravi and H. Mao and C. Rolland and L. Gustafson and T. Xiao and S. Whitehead and A. C. Berg and W. Y. Lo and P. Doll \'a r and R. Girshick. Segment Anything. 2023

    work page 2023

  53. [53]

    International Conference on Learning Representations , year=

    Ravi, Nikhila and Gabeur, Valentin and Hu, Yuan-Ting and Hu, Ronghang and Ryali, Chaitanya and Ma, Tengyu and Khedr, Haitham and R. International Conference on Learning Representations , year=

    work page

  54. [54]

    Li and J

    H. Li and J. Yang and Y. Xu and R. Wang. A Level Set Annotation Framework With Single-Point Supervision for Infrared Small Target Detection. IEEE Signal Processing Letters. 2024. doi:10.1109/LSP.2024.3356411

    work page doi:10.1109/lsp.2024.3356411 2024

  55. [55]

    Liu and Z

    R. Liu and Z. Liu and W. Yao and S. Zeng and J. Zhang. Moreau envelope for nonconvex Bi-level optimization: a single-loop and Hessian-free solution strategy. Proceedings of the 41st International Conference on Machine Learning. 2024

    work page 2024

  56. [56]

    Ying and L

    X. Ying and L. Liu and Y. Wang and R. Li and N. Chen and Z. Lin and W. Sheng and S. Zhou. Mapping Degeneration Meets Label Evolution: Learning Infrared Small Target Detection With Single Point Supervision. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2023

    work page 2023

  57. [57]

    Zhang and R

    M. Zhang and R. Zhang and Y. Yang and H. Bai and J. Zhang and J. Guo. ISNet : Shape matters for infrared small target detection. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022

    work page 2022

  58. [58]

    Zhang and K

    M. Zhang and K. Yue and J. Zhang and Y. Li and X. Gao. Exploring feature compensation and cross-level correlation for infrared small target detection. Proceedings of the 30th ACM International Conference on Multimedia. 2022

    work page 2022

  59. [59]

    Tomasi and R

    C. Tomasi and R. Manduchi. Bilateral filtering for gray and color images. Sixth International Conference on Computer Vision (IEEE Cat. No. 98CH36271). 1998

    work page 1998

  60. [60]

    Zhang and C

    M. Zhang and C. Zhang and Q. Zhang and Y. Li and X. Gao and J. Zhang. Unleashing the power of generic segmentation model: A simple baseline for infrared small target detection. Proceedings of the 32nd ACM International Conference on Multimedia. 2024

    work page 2024

  61. [61]

    Yu and J

    C. Yu and J. Zhao and Y. Liu and S. Zhao and Y. Dai and X. Yue. From easy to hard: Progressive active learning framework for infrared small target detection with single point supervision. Proceedings of the IEEE/CVF International Conference on Computer Vision. 2025

    work page 2025

  62. [62]

    Li and Y

    B. Li and Y. Wang and L. Wang and F. Zhang and T. Liu and Z. Lin and W. An and Y. Guo. Monte Carlo linear clustering with single-point supervision is enough for infrared small target detection. Proceedings of the IEEE/CVF International Conference on Computer Vision. 2023

    work page 2023

  63. [63]

    Zhang and W

    M. Zhang and W. Shang and F. Gao and Q. Zhang and F. Lu and J. Zhang. Semi-supervised Infrared Small Target Detection with Thermodynamic-Inspired Uneven Perturbation and Confidence Adaptation. Proceedings of the AAAI Conference on Artificial Intelligence. 2025

    work page 2025

  64. [64]

    Dai and Y

    Y. Dai and Y. Wu and F. Zhou and K. Barnard. Asymmetric contextual modulation for infrared small target detection. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2021

    work page 2021

  65. [65]

    Ryu and S

    J. Ryu and S. Kim. Small infrared target detection by data-driven proposal and deep learning-based classification. Infrared Technology and Applications XLIV. 2018

    work page 2018

  66. [66]

    Zhang and Y

    M. Zhang and Y. Wang and J. Guo and Y. Li and X. Gao and J. Zhang. IRSAM : Advancing segment anything model for infrared small target detection. European Conference on Computer Vision. 2024

    work page 2024

  67. [67]

    Chen and Z

    T. Chen and Z. Tan and Q. Chu and Y. Wu and B. Liu and N. Yu. TCI-Former : Thermal conduction-inspired transformer for infrared small target detection. Proceedings of the AAAI Conference on Artificial Intelligence. 2024

    work page 2024

  68. [68]

    Wang and J

    Y. Wang and J. Zhao and Z. Fan and X. Zhang and X. Wu and Y. Zhang and L. Jin and X. Li and G. Wang and M. Jia and others. JTD-UAV : MLLM-enhanced Joint Tracking and Description Framework for anti-UAV Systems. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2025

    work page 2025

  69. [69]

    Zhang and X

    M. Zhang and X. Li and F. Gao and J. Guo and X. Gao and J. Zhang. SAIST : Segment Any Infrared Small Target Model Guided by Contrastive Language-Image Pretraining. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2025

    work page 2025

  70. [70]

    Huang and S

    F. Huang and S. Zheng and Z. Qiu and H. Liu and H. Bai and L. Chen. Text-IRSTD : Leveraging Semantic Text to Promote Infrared Small Target Detection in Complex Scenes. Proceedings of the IEEE/CVF International Conference on Computer Vision. 2025

    work page 2025

  71. [71]

    Wang and L

    H. Wang and L. Zhou and L. Wang. Miss detection vs. false alarm: Adversarial learning for small object segmentation in infrared images. Proceedings of the IEEE/CVF International Conference on Computer Vision. 2019

    work page 2019

  72. [72]

    Guo and Q

    C. Guo and Q. Ma and L. Zhang. Spatio-temporal saliency detection using phase spectrum of quaternion fourier transform. 2008 IEEE Conference on Computer Vision and Pattern Recognition. 2008

    work page 2008

  73. [73]

    Metzger and R

    N. Metzger and R. C. Daudt and K. Schindler. Guided depth super-resolution by deep anisotropic diffusion. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2023

    work page 2023

  74. [74]

    Chen and Q

    T. Chen and Q. Chu and B. Liu and N. Yu. Fluid Dynamics-Inspired Network for Infrared Small Target Detection. International Joint Conference on Artificial Intelligence. 2023

    work page 2023

  75. [75]

    Yang and S

    J. Yang and S. Liu and J. Wu and X. Su and N. Hai and X. Huang. Pinwheel-shaped convolution and scale-based dynamic loss for infrared small target detection. Proceedings of the AAAI Conference on Artificial Intelligence. 2025

    work page 2025

  76. [76]

    Liu and R

    Q. Liu and R. Liu and B. Zheng and H. Wang and Y. Fu. Infrared small target detection with scale and location sensitivity. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2024

    work page 2024

  77. [77]

    Borgnakke and R

    C. Borgnakke and R. E. Sonntag. Fundamentals of thermodynamics. 2020

    work page 2020

  78. [78]

    International Joint Conference on Artificial Intelligence

    Bi-level dynamic learning for jointly multi-modality image fusion and beyond , author=. International Joint Conference on Artificial Intelligence. 2023

    work page 2023

  79. [79]

    Liu, Zhu and Wang, Zijun and Liu, Jinyuan and Meng, Fanqi and Ma, Long and Liu, Risheng , booktitle=

    work page

  80. [80]

    Liu, Zhu and Liu, Jinyuan and Zhang, Benzhuang and Ma, Long and Fan, Xin and Liu, Risheng , booktitle=

    work page