pith. sign in

arxiv: 2606.22353 · v1 · pith:CHNOGDKKnew · submitted 2026-06-21 · 💻 cs.CV

Interest Entanglement: The Hidden Barrier to Blind Super-Resolution Optimization

Junxiong Lin , Xinji Mai , Qianyu Guo , Haoran Wang , Zeng Tao , Xuan Tong , Ivy Pan , Wenqiang Zhang This is my paper

Pith reviewed 2026-06-26 10:50 UTC · model grok-4.3

classification 💻 cs.CV
keywords image super-resolutioninterest entanglementmulti-objective optimizationshared feature representationregression lossperceptual lossInfoSqueeze modulefrequency-domain conflict
0
0 comments X

The pith

Decoupling regression and perceptual objectives via shared feature representations resolves interest entanglement in super-resolution.

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

The paper identifies interest entanglement as the barrier to balancing fidelity and perceptual quality in image super-resolution, arising from an inherent frequency-domain conflict between regression and perception objectives. Existing methods that adjust loss weights overlook this entanglement. The authors propose the Shared-Feature-Representation based Super-Resolution framework (SFR) to decouple the learning processes of the two objectives. This decoupling lets the model find a common optimization direction. The InfoSqueeze module aids the process by filtering redundant information through dimensionality reduction and expansion to create consistent feature spaces, with experiments on five datasets supporting the resulting balance.

Core claim

The paper claims that the interest entanglement problem, arising from the inherent frequency-domain conflict between regression and perceptual objectives in SR tasks, can be addressed by the Shared-Feature-Representation based Super-Resolution framework (SFR), which decouples the learning processes of different optimization objectives to explore a common direction and achieve effective balance, aided by the InfoSqueeze module for feature consistency.

What carries the argument

The Shared-Feature-Representation based Super-Resolution framework (SFR) with the InfoSqueeze module, which decouples learning of regression and perceptual objectives by transforming shared features into a consistent space.

If this is right

  • The model explores a common optimization direction for both regression and perceptual goals.
  • An effective balance between fidelity and perceptual quality is achieved without relying on loss weight adjustments alone.
  • Redundant information is filtered and features are transformed into a consistent space through the InfoSqueeze module.
  • Quantitative and qualitative results improve across five representative datasets compared to prior methods.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar decoupling via shared representations could apply to other multi-objective optimization problems in vision tasks such as denoising or enhancement.
  • The frequency-domain analysis of the conflict suggests potential for new loss designs that minimize entanglement from the start.
  • Extending the approach to blind super-resolution may require additional handling of unknown degradations to maintain the shared feature consistency.

Load-bearing premise

The frequency-domain conflict between regression and perceptual objectives can be resolved by decoupling via shared feature representations and the InfoSqueeze module.

What would settle it

An experiment showing that SFR yields no measurable improvement in balance metrics over standard weighted-loss baselines on the five representative datasets would falsify the claim that decoupling resolves the entanglement.

Figures

Figures reproduced from arXiv: 2606.22353 by Haoran Wang, Ivy Pan, Junxiong Lin, Qianyu Guo, Wenqiang Zhang, Xinji Mai, Xuan Tong, Zeng Tao.

Figure 1
Figure 1. Figure 1: Frequency domain energy distribution of High-Resolution (HR) images, Low-Resolution (LR) images, and Super [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The Shared-Feature-Representation based Super-Resolution framework (SFR). SFR consists of three branches: the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Visual comparisons of several representative methods on examples of the Urban100 dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Visualization of ablation study on different [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
read the original abstract

Fidelity and perceptual quality are two inherently competing and conflicting objectives in the image super-resolution (SR) task. Different loss functions focus on these objectives to varying extents. Regression losses enhance the model's fidelity but lack sufficient attention to high-frequency details, resulting in a loss of fine details. In contrast, perception losses improve the model's visual quality but may introduce undesirable artifacts. Balancing these two optimization goals can be viewed as a Multi-Objective Optimization problem. Existing methods are limited to cautiously adjusting weight parameters between these losses, overlooking the underlying Interest Entanglement problem. To address this problem, we explore the inherent frequency-domain conflict between the regression objective and the perceptual objective, and analyze the causes of Interest Entanglement in SR tasks. According to our findings, we propose the Shared-Feature-Representation based Super-Resolution framework (SFR), which decouples the learning process of different optimization objectives, allowing the model to explore a common optimization direction for both goals and achieve an effective balance between them. To better leverage shared features, we also proposed the InfoSqueeze module, which filters redundant information through a dimensionality reduction and expansion process, effectively transforming features into a consistent space. Quantitative and qualitative experiments across five representative datasets affirm the superiority of SFR.

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 / 1 minor

Summary. The paper identifies an 'Interest Entanglement' problem in image super-resolution arising from inherent frequency-domain conflicts between regression (fidelity) and perceptual losses. It argues that existing methods are limited to loss-weight tuning and proposes the Shared-Feature-Representation based Super-Resolution (SFR) framework together with an InfoSqueeze module that performs dimensionality reduction/expansion to transform features into a consistent space, thereby decoupling the objectives and enabling a common optimization direction. Quantitative and qualitative results on five datasets are said to demonstrate superiority.

Significance. If the SFR framework and InfoSqueeze module can be shown to produce a genuinely common optimization direction rather than a weighted compromise, the work would offer a structural alternative to hyper-parameter tuning for multi-objective SR, with potential impact on other tasks that balance pixel-wise fidelity against perceptual quality.

major comments (2)
  1. [Abstract] Abstract: the central claim that SFR 'decouples the learning process of different optimization objectives' and that InfoSqueeze 'effectively transforming features into a consistent space' is load-bearing yet unsupported by any gradient-flow analysis, frequency-band isolation derivation, or mathematical formulation; without this, it remains possible that both losses continue to back-propagate through the shared representation and produce an entangled compromise rather than a common direction.
  2. [Abstract] Abstract: the assertion of 'superiority' across five datasets is stated without any reported metrics, baselines, error bars, or ablation results, preventing assessment of whether the claimed balance between fidelity and perceptual quality is actually achieved.
minor comments (1)
  1. [Abstract] The term 'Interest Entanglement' is introduced as a novel concept but is not formally defined or distinguished from related ideas in multi-task or conflicting-objective optimization literature.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address the two major comments on the abstract below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that SFR 'decouples the learning process of different optimization objectives' and that InfoSqueeze 'effectively transforming features into a consistent space' is load-bearing yet unsupported by any gradient-flow analysis, frequency-band isolation derivation, or mathematical formulation; without this, it remains possible that both losses continue to back-propagate through the shared representation and produce an entangled compromise rather than a common direction.

    Authors: We acknowledge that the abstract states the decoupling claim without an explicit gradient-flow derivation. Section 3 of the manuscript provides the frequency-domain analysis of Interest Entanglement and motivates the SFR framework via the shared representation. To strengthen the argument that this yields a common optimization direction rather than a compromise, we will add a concise mathematical formulation (including a sketch of the gradient flow through the shared features) in the revised abstract and Section 3. revision: yes

  2. Referee: [Abstract] Abstract: the assertion of 'superiority' across five datasets is stated without any reported metrics, baselines, error bars, or ablation results, preventing assessment of whether the claimed balance between fidelity and perceptual quality is actually achieved.

    Authors: The abstract summarizes the experimental outcome; the full results (PSNR/SSIM for fidelity, LPIPS/FID for perception, comparisons against ESRGAN, Real-ESRGAN and other baselines, plus ablations) appear in Section 4 with tables and figures on the five datasets. We will revise the abstract to include the key quantitative improvements and note that error bars are reported from multiple runs. revision: yes

Circularity Check

0 steps flagged

No circularity: independent framework proposal

full rationale

The paper identifies the interest entanglement problem in SR, attributes it to frequency-domain conflict between regression and perceptual losses, and proposes the SFR framework plus InfoSqueeze module as a new architectural solution to decouple objectives. No equations, fitted parameters, or derivations are shown that reduce the claimed decoupling or common optimization direction to the inputs by construction. No self-citations are invoked as load-bearing uniqueness theorems, no ansatzes are smuggled, and no known results are merely renamed. The derivation chain is self-contained as an independent proposal rather than a redefinition or statistical forcing of outputs from inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 1 invented entities

The central claim rests on the newly introduced concept of Interest Entanglement and the unverified effectiveness of the SFR framework and InfoSqueeze module for decoupling objectives.

invented entities (1)
  • Interest Entanglement no independent evidence
    purpose: To name the hidden barrier arising from frequency-domain conflict between fidelity and perceptual objectives
    New term coined in the abstract to frame the optimization issue.

pith-pipeline@v0.9.1-grok · 5766 in / 1237 out tokens · 38677 ms · 2026-06-26T10:50:27.290329+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

119 extracted references · 25 canonical work pages

  1. [1]

    IEEE Transactions on Acoustics, Speech, and Signal Processing , author=

    Cubic splines for image interpolation and digital filtering , volume=. IEEE Transactions on Acoustics, Speech, and Signal Processing , author=. 1978 , month=. doi:10.1109/tassp.1978.1163154 , number=

    work page doi:10.1109/tassp.1978.1163154 1978

  2. [2]

    Image Interpolation and Resampling , url=

    Thévenaz, Philippe and Blu, Thierry and Unser, Michael , year=. Image Interpolation and Resampling , url=. doi:10.1016/b978-012077790-7/50030-8 , booktitle=

    work page doi:10.1016/b978-012077790-7/50030-8

  3. [3]

    The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Gu, Jinjin and Lu, Hannan and Zuo, Wangmeng and Dong, Chao , title =. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , month =

  4. [4]

    Advances in Neural Information Processing Systems , volume=

    Blind super-resolution kernel estimation using an internal-gan , author=. Advances in Neural Information Processing Systems , volume=

  5. [5]

    IEEE International Conference on Computer Vision , year=

    Mutual Affine Network for Spatially Variant Kernel Estimation in Blind Image Super-Resolution , author=. IEEE International Conference on Computer Vision , year=

  6. [6]

    arXiv preprint arXiv:2108.10257 , year=

    SwinIR: Image Restoration Using Swin Transformer , author=. arXiv preprint arXiv:2108.10257 , year=

    arXiv

  7. [7]

    CVPR , year=

    Restormer: Efficient Transformer for High-Resolution Image Restoration , author=. CVPR , year=

  8. [8]

    arXiv preprint arXiv:2309.05239 , year=

    HAT: Hybrid Attention Transformer for Image Restoration , author=. arXiv preprint arXiv:2309.05239 , year=

    arXiv

  9. [9]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Chen, Xiangyu and Wang, Xintao and Zhou, Jiantao and Qiao, Yu and Dong, Chao , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2023 , pages =

    2023

  10. [10]

    ICCV , year=

    Dual Aggregation Transformer for Image Super-Resolution , author=. ICCV , year=

  11. [11]

    arXiv preprint arXiv:2305.07015 , year =

    Wang, Jianyi and Yue, Zongsheng and Zhou, Shangchen and Chan, Kelvin CK and Loy, Chen Change , title =. arXiv preprint arXiv:2305.07015 , year =

    arXiv

  12. [12]

    Advances in Neural Information Processing Systems , volume=

    Resshift: Efficient diffusion model for image super-resolution by residual shifting , author=. Advances in Neural Information Processing Systems , volume=

  13. [13]

    The Eleventh International Conference on Learning Representations , year=

    Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model , author=. The Eleventh International Conference on Learning Representations , year=

  14. [14]

    2021 , eprint=

    High-Resolution Image Synthesis with Latent Diffusion Models , author=. 2021 , eprint=

    2021

  15. [15]

    Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning , journal=

    Gal, Yarin and Ghahramani, Zoubin , year=. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning , journal=

  16. [16]

    Uncertainty-Aware Variate Decomposition for Self-supervised Blind Image Deblurring , booktitle =

    Runhua Jiang and Yahong Han , editor =. Uncertainty-Aware Variate Decomposition for Self-supervised Blind Image Deblurring , booktitle =. 2023 , url =. doi:10.1145/3581783.3612535 , timestamp =

    work page doi:10.1145/3581783.3612535 2023

  17. [17]

    Handling Label Uncertainty for Camera Incremental Person Re-Identification , booktitle =

    Zexian Yang and Dayan Wu and Wanqian Zhang and Bo Li and Weiping Wang , editor =. Handling Label Uncertainty for Camera Incremental Person Re-Identification , booktitle =. 2023 , url =. doi:10.1145/3581783.3612294 , timestamp =

    work page doi:10.1145/3581783.3612294 2023

  18. [18]

    In: Proceedings of the 31st ACM International Conference on Multimedia

    Hanwei Liu and Huiling Cai and Qingcheng Lin and Xuefeng Li and Hui Xiao , editor =. Learning from More: Combating Uncertainty Cross-multidomain for Facial Expression Recognition , booktitle =. 2023 , url =. doi:10.1145/3581783.3611702 , timestamp =

    work page doi:10.1145/3581783.3611702 2023

  19. [19]

    Semi-Supervised Convolutional Vision Transformer with Bi-Level Uncertainty Estimation for Medical Image Segmentation , booktitle =

    Huimin Huang and Yawen Huang and Shiao Xie and Lanfen Lin and Ruofeng Tong and Yen. Semi-Supervised Convolutional Vision Transformer with Bi-Level Uncertainty Estimation for Medical Image Segmentation , booktitle =. 2023 , url =. doi:10.1145/3581783.3611821 , timestamp =

    work page doi:10.1145/3581783.3611821 2023

  20. [20]

    Orthogonal Uncertainty Representation of Data Manifold for Robust Long-Tailed Learning , booktitle =

    Yanbiao Ma and Licheng Jiao and Fang Liu and Shuyuan Yang and Xu Liu and Lingling Li , editor =. Orthogonal Uncertainty Representation of Data Manifold for Robust Long-Tailed Learning , booktitle =. 2023 , url =. doi:10.1145/3581783.3611698 , timestamp =

    work page doi:10.1145/3581783.3611698 2023

  21. [21]

    Uncertainty-Guided End-to-End Audio-Visual Speaker Diarization for Far-Field Recordings , booktitle =

    Chenyu Yang and Mengxi Chen and Yanfeng Wang and Yu Wang , editor =. Uncertainty-Guided End-to-End Audio-Visual Speaker Diarization for Far-Field Recordings , booktitle =. 2023 , url =. doi:10.1145/3581783.3612424 , timestamp =

    work page doi:10.1145/3581783.3612424 2023

  22. [22]

    Symmetric Uncertainty-Aware Feature Transmission for Depth Super-Resolution , booktitle =

    Wuxuan Shi and Mang Ye and Bo Du , editor =. Symmetric Uncertainty-Aware Feature Transmission for Depth Super-Resolution , booktitle =. 2022 , url =. doi:10.1145/3503161.3547873 , timestamp =

    work page doi:10.1145/3503161.3547873 2022

  23. [23]

    Pixel Exclusion: Uncertainty-aware Boundary Discovery for Active Cross-Domain Semantic Segmentation , booktitle =

    Fuming You and Jingjing Li and Zhi Chen and Lei Zhu , editor =. Pixel Exclusion: Uncertainty-aware Boundary Discovery for Active Cross-Domain Semantic Segmentation , booktitle =. 2022 , url =. doi:10.1145/3503161.3548079 , timestamp =

    work page doi:10.1145/3503161.3548079 2022

  24. [24]

    Pixel-Level Anomaly Detection via Uncertainty-aware Prototypical Transformer , booktitle =

    Chao Huang and Chengliang Liu and Zheng Zhang and Zhihao Wu and Jie Wen and Qiuping Jiang and Yong Xu , editor =. Pixel-Level Anomaly Detection via Uncertainty-aware Prototypical Transformer , booktitle =. 2022 , url =. doi:10.1145/3503161.3548082 , timestamp =

    work page doi:10.1145/3503161.3548082 2022

  25. [25]

    Uncertainty-Aware Semi-Supervised Learning of 3D Face Rigging from Single Image , booktitle =

    Yong Zhao and Haifeng Chen and Hichem Sahli and Ke Lu and Dongmei Jiang , editor =. Uncertainty-Aware Semi-Supervised Learning of 3D Face Rigging from Single Image , booktitle =. 2022 , url =. doi:10.1145/3503161.3548285 , timestamp =

    work page doi:10.1145/3503161.3548285 2022

  26. [26]

    Uncertainty-Aware 3D Human Pose Estimation from Monocular Video , booktitle =

    Jinlu Zhang and Yujin Chen and Zhigang Tu , editor =. Uncertainty-Aware 3D Human Pose Estimation from Monocular Video , booktitle =. 2022 , url =. doi:10.1145/3503161.3547773 , timestamp =

    work page doi:10.1145/3503161.3547773 2022

  27. [27]

    Modeling the Uncertainty for Self-supervised 3D Skeleton Action Representation Learning , booktitle =

    Yukun Su and Guosheng Lin and Ruizhou Sun and Yun Hao and Qingyao Wu , editor =. Modeling the Uncertainty for Self-supervised 3D Skeleton Action Representation Learning , booktitle =. 2021 , url =. doi:10.1145/3474085.3475248 , timestamp =

    work page doi:10.1145/3474085.3475248 2021

  28. [28]

    Handling Difficult Labels for Multi-label Image Classification via Uncertainty Distillation , booktitle =

    Liangchen Song and Jialian Wu and Ming Yang and Qian Zhang and Yuan Li and Junsong Yuan , editor =. Handling Difficult Labels for Multi-label Image Classification via Uncertainty Distillation , booktitle =. 2021 , url =. doi:10.1145/3474085.3475406 , timestamp =

    work page doi:10.1145/3474085.3475406 2021

  29. [29]

    Uncertainty-aware Cross-dataset Facial Expression Recognition via Regularized Conditional Alignment , booktitle =

    Linyi Zhou and Xijian Fan and Yingjie Ma and Tardi Tjahjadi and Qiaolin Ye , editor =. Uncertainty-aware Cross-dataset Facial Expression Recognition via Regularized Conditional Alignment , booktitle =. 2020 , url =. doi:10.1145/3394171.3413515 , timestamp =

    work page doi:10.1145/3394171.3413515 2020

  30. [30]

    , Kendall , A

    Badrinarayanan, Vijay and Kendall, Alex and Cipolla, Roberto , year=. SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation , url=. doi:10.1109/tpami.2016.2644615 , journal=

    work page doi:10.1109/tpami.2016.2644615 2016

  31. [31]

    arXiv preprint arXiv:2303.17783 , year=

    SOSR: Source-free image super-resolution with wavelet augmentation transformer , author=. arXiv preprint arXiv:2303.17783 , year=

    arXiv

  32. [32]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Spectral Bayesian uncertainty for image super-resolution , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  33. [33]

    Advances in Neural Information Processing Systems , volume=

    Uncertainty-driven loss for single image super-resolution , author=. Advances in Neural Information Processing Systems , volume=

  34. [34]

    Artificial Intelligence Review , volume=

    A survey of uncertainty in deep neural networks , author=. Artificial Intelligence Review , volume=. 2023 , publisher=

    2023

  35. [35]

    GRAM: Gradient Rescaling Attention Model for Data Uncertainty Estimation in Single Image Super Resolution , journal=

    Lee, Chang-Woo and Ki-Seok, Chung , year=. GRAM: Gradient Rescaling Attention Model for Data Uncertainty Estimation in Single Image Super Resolution , journal=

  36. [36]

    Fast Bayesian Uncertainty Estimation of Batch Normalized Single Image Super-Resolution Network

    Kar, Aupendu and Biswas, PrabirKumar , year=. Fast Bayesian Uncertainty Estimation of Batch Normalized Single Image Super-Resolution Network. , journal=

  37. [37]

    The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , month =

    Lim, Bee and Son, Sanghyun and Kim, Heewon and Nah, Seungjun and Lee, Kyoung Mu , title =. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , month =

  38. [38]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Deep constrained least squares for blind image super-resolution , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  39. [39]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Blind image super-resolution with elaborate degradation modeling on noise and kernel , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  40. [40]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Learning Correction Filter via Degradation-Adaptive Regression for Blind Single Image Super-Resolution , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  41. [41]

    Kernel Dimension Matters: To Activate Available Kernels for Real-time Video Super-Resolution , booktitle =

    Shuo Jin and Meiqin Liu and Chao Yao and Chunyu Lin and Yao Zhao , editor =. Kernel Dimension Matters: To Activate Available Kernels for Real-time Video Super-Resolution , booktitle =. 2023 , url =. doi:10.1145/3581783.3611908 , timestamp =

    work page doi:10.1145/3581783.3611908 2023

  42. [42]

    Blind Image Super-resolution with Rich Texture-Aware Codebook , booktitle =

    Rui Qin and Ming Sun and Fangyuan Zhang and Xing Wen and Bin Wang , editor =. Blind Image Super-resolution with Rich Texture-Aware Codebook , booktitle =. 2023 , url =. doi:10.1145/3581783.3611917 , timestamp =

    work page doi:10.1145/3581783.3611917 2023

  43. [43]

    Blind Image Super-resolution with Rich Texture-Aware Codebook , booktitle =

    Wenyu Zhang and Xin Deng and Baojun Jia and Xingtong Yu and Yifan Chen and Jin Ma and Qing Ding and Xinming Zhang , editor =. Pixel Adapter:. Proceedings of the 31st. 2023 , url =. doi:10.1145/3581783.3611913 , timestamp =

    work page doi:10.1145/3581783.3611913 2023

  44. [44]

    Proceedings of the 31st

    Boyang Wang and Yan Wang and Qing Zhao and Junxiong Lin and Zeng Tao and Pinxue Guo and Zhaoyu Chen and Kaixun Jiang and Shaoqi Yan and Shuyong Gao and Wenqiang Zhang , title =. Proceedings of the 31st. 2023 , url =

    2023

  45. [45]

    Proceedings of the 31st

    Zhou Zhou and Jiahao Chao and Jiali Gong and Hongfan Gao and Zhenbing Zeng and Zhengfeng Yang , title =. Proceedings of the 31st. 2023 , url =

    2023

  46. [46]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Suppressing uncertainties for large-scale facial expression recognition , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  47. [47]

    and Coll, B

    Buades, A. and Coll, B. and Morel, J.-M. , year=. A Non-Local Algorithm for Image Denoising , url=. doi:10.1109/cvpr.2005.38 , booktitle=

    work page doi:10.1109/cvpr.2005.38 2005

  48. [48]

    Korean journal of anesthesiology , volume=

    Central limit theorem: the cornerstone of modern statistics , author=. Korean journal of anesthesiology , volume=. 2017 , publisher=

    2017

  49. [49]

    and Welling, Max , year=

    Kingma, DiederikP. and Welling, Max , year=. Auto-Encoding Variational Bayes , journal=

  50. [50]

    IEEE Transactions on pattern analysis and machine intelligence , number=

    Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images , author=. IEEE Transactions on pattern analysis and machine intelligence , number=. 1984 , publisher=

    1984

  51. [51]

    Super-Resolution with Deep Convolutional Sufficient Statistics , journal=

    Bruna, Joan and Sprechmann, Pablo and LeCun, Yann , year=. Super-Resolution with Deep Convolutional Sufficient Statistics , journal=

  52. [52]

    arXiv preprint arXiv:2201.10084 , year=

    Revisiting L1 loss in super-resolution: a probabilistic view and beyond , author=. arXiv preprint arXiv:2201.10084 , year=

    arXiv

  53. [53]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Swin transformer: Hierarchical vision transformer using shifted windows , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  54. [54]

    2020 , month=

    Unified Dynamic Convolutional Network for Super-Resolution with Variational Degradations , author=. 2020 , month=

    2020

  55. [55]

    Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XXII 16 , pages=

    Latticenet: Towards lightweight image super-resolution with lattice block , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XXII 16 , pages=. 2020 , organization=

    2020

  56. [56]

    arXiv preprint arXiv:2105.11683 , year=

    Towards compact single image super-resolution via contrastive self-distillation , author=. arXiv preprint arXiv:2105.11683 , year=

    arXiv

  57. [57]

    The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , month =

    Agustsson, Eirikur and Timofte, Radu , title =. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , month =

  58. [58]

    and Fowlkes, C

    Arbelaez, Pablo and Maire, Michael and Fowlkes, Charless and Malik, Jitendra , title =. IEEE Trans. Pattern Anal. Mach. Intell. , issue_date =. 2011 , issn =. doi:10.1109/TPAMI.2010.161 , acmid =

    work page doi:10.1109/tpami.2010.161 2011

  59. [59]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

    Single Image Super-Resolution From Transformed Self-Exemplars , Author =. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

  60. [60]

    Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part VIII 16 , pages=

    Component divide-and-conquer for real-world image super-resolution , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part VIII 16 , pages=. 2020 , organization=

    2020

  61. [61]

    Proceedings of the IEEE International Conference on Computer Vision , pages=

    DSLR-Quality Photos on Mobile Devices with Deep Convolutional Networks , author=. Proceedings of the IEEE International Conference on Computer Vision , pages=

  62. [62]

    IEEE Transactions on Image Processing19(11), 2861–2873 (2010)

    Jianchao Yang and Wright, John and Huang, Thomas S and Yi Ma , year=. Image Super-Resolution Via Sparse Representation , url=. doi:10.1109/tip.2010.2050625 , journal=

    work page doi:10.1109/tip.2010.2050625 2010

  63. [63]

    Advances in Neural Information Processing Systems (NeurIPS) , volume=

    Unfolding the Alternating Optimization for Blind Super Resolution , author=. Advances in Neural Information Processing Systems (NeurIPS) , volume=

  64. [64]

    arXiv preprint arXiv:2311.16512 , year=

    CoSeR: Bridging Image and Language for Cognitive Super-Resolution , author=. arXiv preprint arXiv:2311.16512 , year=

    arXiv

  65. [65]

    IEEE Transactions on Image Processing , year=

    Meta-learning based degradation representation for blind super-resolution , author=. IEEE Transactions on Image Processing , year=

  66. [66]

    arXiv preprint arXiv:2403.04294 , year=

    A3lign-DFER: Pioneering Comprehensive Dynamic Affective Alignment for Dynamic Facial Expression Recognition with CLIP , author=. arXiv preprint arXiv:2403.04294 , year=

    arXiv

  67. [67]

    arXiv preprint arXiv:2405.18769 , year=

    OUS: Scene-Guided Dynamic Facial Expression Recognition , author=. arXiv preprint arXiv:2405.18769 , year=

    arXiv

  68. [68]

    2023 , isbn =

    Tao, Zeng and Wang, Yan and Chen, Zhaoyu and Wang, Boyang and Yan, Shaoqi and Jiang, Kaixun and Gao, Shuyong and Zhang, Wenqiang , title =. 2023 , isbn =. doi:10.1145/3581783.3611972 , booktitle =

    work page doi:10.1145/3581783.3611972 2023

  69. [69]

    Engineering Applications of Artificial Intelligence , volume=

    Transformer Autoencoder for K-means Efficient clustering , author=. Engineering Applications of Artificial Intelligence , volume=. 2024 , publisher=

    2024

  70. [70]

    IEEE Transactions on Image Processing , volume=

    Deep red unfolding network for image restoration , author=. IEEE Transactions on Image Processing , volume=. 2021 , publisher=

    2021

  71. [71]

    arXiv preprint arXiv:2406.16473 , year=

    Seeking Certainty In Uncertainty: Dual-Stage Unified Framework Solving Uncertainty in Dynamic Facial Expression Recognition , author=. arXiv preprint arXiv:2406.16473 , year=

    arXiv

  72. [72]

    IEEE Transactions on Image Processing , volume=

    Two-direction nonlocal model for image denoising , author=. IEEE Transactions on Image Processing , volume=. 2012 , publisher=

    2012

  73. [73]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Query-aware quantization for maximum inner product search , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  74. [74]

    Applied Soft Computing , volume=

    Multi-objective evolving long--short term memory networks with attention for network intrusion detection , author=. Applied Soft Computing , volume=. 2023 , publisher=

    2023

  75. [75]

    IEEE communications surveys & tutorials , volume=

    Software-defined networking (SDN) and distributed denial of service (DDoS) attacks in cloud computing environments: A survey, some research issues, and challenges , author=. IEEE communications surveys & tutorials , volume=. 2015 , publisher=

    2015

  76. [76]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Efficient decision-based black-box patch attacks on video recognition , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  77. [77]

    2024 , eprint=

    From Efficient Multimodal Models to World Models: A Survey , author=. 2024 , eprint=

    2024

  78. [78]

    Pattern Recognition , volume=

    LCSeg-Net: A low-contrast images semantic segmentation model with structural and frequency spectrum information , author=. Pattern Recognition , volume=. 2024 , publisher=

    2024

  79. [79]

    2024 , eprint=

    All rivers run into the sea: Unified Modality Brain-like Emotional Central Mechanism , author=. 2024 , eprint=

    2024

  80. [80]

    2024 , eprint=

    Hi-EF: Benchmarking Emotion Forecasting in Human-interaction , author=. 2024 , eprint=

    2024

Showing first 80 references.