pith. sign in

arxiv: 2604.20594 · v1 · submitted 2026-04-22 · 💻 cs.CV

Physics-Informed Conditional Diffusion for Motion-Robust Retinal Temporal Laser Speckle Contrast Imaging

Qian Chen , Yuehao Chen , Qiang Wang , Lei Zhu , Yanye Lu , Qiushi Ren This is my paper

Pith reviewed 2026-05-10 00:14 UTC · model grok-4.3

classification 💻 cs.CV
keywords retinal laser speckle contrast imagingconditional diffusion modelsmotion correctiontemporal reconstructionphysics-informed reconstructionblood flow imagingfew-frame imaging
0
0 comments X

The pith

A conditional diffusion model guided by a motion-corrected physics prior reconstructs stable retinal temporal laser speckle contrast images from only a few frames.

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

The paper introduces RetinaDiff to overcome the requirement for long speckle sequences in conventional temporal laser speckle contrast imaging of the retina. Phase correlation first registers the short raw sequence to reduce misalignment and yields a corrected physics prior together with a high-quality multiframe reference. A conditional diffusion model then performs the inverse reconstruction by conditioning jointly on the registered sequence and this prior. The result is improved structural continuity and statistical stability relative to direct few-frame computation or standard baselines, and the method holds up even when both the five-frame input and conventional multiframe outputs are badly degraded by motion.

Core claim

RetinaDiff first applies phase correlation registration to stabilize the raw speckle sequence, thereby supplying a motion-corrected physics prior and a high-quality multiframe tLSCI reference; a conditional diffusion model then carries out inverse reconstruction by jointly conditioning on the registered speckle sequence and the corrected prior. On data from an in-house retinal LSCI system this produces images with better structural continuity and statistical stability than direct reconstruction from five frames or representative baselines, and the framework remains effective in a small number of extremely challenging acquisitions where both the direct five-frame input and conventional multia

What carries the argument

Conditional diffusion model that jointly conditions on the registered speckle sequence and the motion-corrected physics prior to perform inverse reconstruction.

If this is right

  • Reconstruction of retinal blood-flow contrast becomes possible from extremely short sequences without severe loss of structural continuity.
  • Statistical stability of the contrast values improves relative to direct few-frame computation even under motion.
  • The method continues to produce usable images in acquisitions where both direct five-frame inputs and conventional long-sequence reconstructions are severely degraded.

Where Pith is reading between the lines

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

  • The same registration-plus-conditional-diffusion pattern could be tested on other motion-sensitive optical modalities such as laser Doppler or OCT angiography.
  • If the physics prior is replaced by an alternative motion estimator the framework might still deliver stable outputs, offering a route to reduce frame count further.
  • Clinical monitoring of retinal hemodynamics could move toward higher temporal resolution in unsedated subjects once the registration step is made fully automatic.

Load-bearing premise

Phase correlation registration produces a sufficiently accurate motion-corrected physics prior so that residual misalignment does not propagate errors into the diffusion reconstruction.

What would settle it

Acquire retinal sequences with known large eye movements, compute long-sequence ground-truth tLSCI, and check whether the five-frame RetinaDiff output matches the ground-truth continuity and variance statistics better than direct five-frame or conventional multiframe methods.

Figures

Figures reproduced from arXiv: 2604.20594 by Lei Zhu, Qian Chen, Qiang Wang, Qiushi Ren, Yanye Lu, Yuehao Chen.

Figure 1
Figure 1. Figure 1: Motivation and core phenomenon of retinal temporal laser speckle contrast imaging (tLSCI) reconstruction that is robust to motion and uses only a few frames. (A) Retinal LSCI acquisition and practical clinical challenges (eye motion artifacts cause inter-frame misalignment and blurring; long acquisition time increases patient discomfort and motion risk). (B) Two paradigms of retinal tLSCI reconstruction: u… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of RetinaDiff. Stage 1 performs stabilization based on phase correlation and generates a high quality multiframe tLSCI map. Stage 2 performs physically informed conditional diffusion reconstruction using the aligned speckle sequence of a few frames and the physics prior. temporal inconsistency caused by involuntary eye motion. [19,31] Temporal laser speckle contrast is then computed on the aligned… view at source ↗
Figure 3
Figure 3. Figure 3: Representative results under stable acquisition conditions. The columns correspond to Direct-5f, U-Net[32] , GAN[33] , RetinaDiff, and the HQ Reference reconstructed from the long registered sequence. Insets A and B show enlarged regions of interest with color-coded annotations: red arrows indicate fine small vessels that are clearly resolved in RetinaDiff but absent or fragmented in competing methods; blu… view at source ↗
Figure 4
Figure 4. Figure 4: Visual comparison under challenging nonstationary acquisition conditions. For each case, the images are arranged from left to right as Direct 5-f, Degraded Multi-f, and RetinaDiff. (A) Representative challenging cases, where the direct 5-frame reconstruction remains noisy but still preserves recoverable vascular cues, whereas the conventional multiframe reconstruction is degraded by temporal integration ov… view at source ↗
Figure 5
Figure 5. Figure 5: Ablation analysis of RetinaDiff under stable acquisition conditions. From left to right, the columns show Direct-5f, Prior-only, Raw-only, RetinaDiff, and the HQ Reference. Insets A and B highlight the respective effects of direct statistics from a few frames, reconstruction from the physics prior only, conditioning on the raw input only, and fusion by the full model on vessel continuity, background suppre… view at source ↗
read the original abstract

Retinal laser speckle contrast imaging (LSCI) is a noninvasive optical modality for monitoring retinal blood flow dynamics. However, conventional temporal LSCI (tLSCI) reconstruction relies on sufficiently long speckle sequences to obtain stable temporal statistics, which makes it vulnerable to acquisition disturbances and limits effective temporal resolution. A physically informed reconstruction framework, termed RetinaDiff (Retinal Diffusion Model), is proposed for retinal tLSCI that is robust to motion and requires only a few frames. In RetinaDiff, registration based on phase correlation is first applied to stabilize the raw speckle sequence before contrast computation, reducing interframe misalignment so that fluctuations at each pixel primarily reflect true flow dynamics. This step provides a physics prior corrected for motion and a high quality multiframe tLSCI reference. Next, guided by the physics prior, a conditional diffusion model performs inverse reconstruction by jointly conditioning on the registered speckle sequence and the corrected prior. Experiments on data acquired with a retinal LSCI system developed in house show improved structural continuity and statistical stability compared with direct reconstruction from few frames and representative baselines. The framework also remains effective in a small number of extremely challenging cases, where both the direct 5-frame input and the conventional multiframe reconstruction are severely degraded. Overall, this work provides a practical and physically grounded route for reliable retinal tLSCI reconstruction from extremely limited frames. The source code and model weights will be publicly available at https://github.com/QianChen113/RetinaDiff.

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

3 major / 1 minor

Summary. The manuscript introduces RetinaDiff, a conditional diffusion model for reconstructing temporal laser speckle contrast images (tLSCI) of the retina from a small number of frames. The method first applies phase-correlation-based registration to the raw speckle sequence to generate a motion-corrected physics prior and a high-quality multiframe reference. A conditional diffusion model is then used to perform inverse reconstruction guided by this prior, conditioning on both the registered sequence and the corrected prior. Experiments on in-house retinal LSCI data demonstrate improved structural continuity and statistical stability compared to direct few-frame reconstruction and baselines, with effectiveness maintained in challenging motion cases.

Significance. If the central claims hold under quantitative scrutiny, the work offers a practical route to high-temporal-resolution retinal blood-flow imaging that tolerates motion and short acquisitions, which could broaden clinical applicability of LSCI. The planned public release of code and model weights is a clear strength for reproducibility.

major comments (3)
  1. Abstract: the central claims of 'improved structural continuity and statistical stability' and effectiveness 'in a small number of extremely challenging cases' are stated only qualitatively; no numerical metrics, dataset sizes, error bars, statistical tests, or ablation results are supplied, leaving the magnitude and reliability of the improvement unquantified.
  2. Methods (registration and prior generation): phase-correlation registration is presented as yielding a reliable motion-corrected physics prior that, together with the multiframe reference, forms the conditioning signal for the diffusion model. No residual-alignment error maps, pixel-wise misalignment statistics, or sensitivity analysis to non-rigid retinal deformations (saccades, vessel pulsation) are provided, yet this assumption is load-bearing for the claim that the diffusion step adds value beyond the registered prior.
  3. Experiments: no ablation isolating the effect of registration quality on final contrast statistics, no comparison against non-rigid registration alternatives, and no propagation analysis of residual misalignment into the conditional diffusion output are described. Without these, it remains unclear whether observed gains stem from the learned inverse step or simply from the quality of the physics prior.
minor comments (1)
  1. Abstract: the phrase 'a small number of extremely challenging cases' is imprecise; stating the exact count and the specific failure modes of the baselines would improve clarity.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We sincerely thank the referee for the constructive and detailed feedback. We address each major comment point by point below. We have revised the manuscript to incorporate quantitative metrics, additional validation analyses, and clarifications as suggested, which we believe strengthen the presentation of our work.

read point-by-point responses

  1. Referee: Abstract: the central claims of 'improved structural continuity and statistical stability' and effectiveness 'in a small number of extremely challenging cases' are stated only qualitatively; no numerical metrics, dataset sizes, error bars, statistical tests, or ablation results are supplied, leaving the magnitude and reliability of the improvement unquantified.

    Authors: We agree that the abstract would be strengthened by quantitative support. In the revised version, we have updated the abstract to report key numerical results from our experiments, including average SSIM and CNR improvements with standard deviations over the direct few-frame and baseline methods, the number of retinal sequences in the dataset, and references to statistical tests (e.g., paired t-tests with p-values). We also briefly note the ablation studies performed. These additions quantify the claims while preserving the abstract's length and focus. revision: yes

  2. Referee: Methods (registration and prior generation): phase-correlation registration is presented as yielding a reliable motion-corrected physics prior that, together with the multiframe reference, forms the conditioning signal for the diffusion model. No residual-alignment error maps, pixel-wise misalignment statistics, or sensitivity analysis to non-rigid retinal deformations (saccades, vessel pulsation) are provided, yet this assumption is load-bearing for the claim that the diffusion step adds value beyond the registered prior.

    Authors: We acknowledge the importance of validating the registration step more explicitly. Phase correlation effectively handles the dominant translational motions typical in retinal LSCI acquisitions. To address the concern, the revised Methods section now includes residual-alignment error maps, pixel-wise misalignment statistics across the dataset, and a sensitivity analysis to simulated non-rigid deformations (small saccades and vessel pulsation). These additions show that residual errors are limited and that the conditional diffusion model further refines the output beyond the physics prior alone. revision: yes

  3. Referee: Experiments: no ablation isolating the effect of registration quality on final contrast statistics, no comparison against non-rigid registration alternatives, and no propagation analysis of residual misalignment into the conditional diffusion output are described. Without these, it remains unclear whether observed gains stem from the learned inverse step or simply from the quality of the physics prior.

    Authors: We agree that targeted ablations would clarify the contribution of each component. The revised Experiments section now includes: (1) an ablation varying registration quality (with/without registration and different alignment strengths) and its direct impact on final contrast statistics; (2) comparisons against non-rigid registration alternatives such as optical-flow-based methods; and (3) a propagation analysis with visualizations showing how residual misalignment affects the diffusion output. These results confirm that the observed improvements arise from the learned inverse reconstruction guided by the physics prior, not solely from the prior itself. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation chain is self-contained

full rationale

The paper describes a sequential pipeline in which phase-correlation registration is first applied to the raw speckle sequence to produce an independent motion-corrected physics prior and multiframe tLSCI reference; a conditional diffusion model then performs reconstruction by conditioning on the registered sequence and this prior. No equations reduce the final contrast image to a fitted parameter defined from the same data, no self-definitional loop exists between prior and output, and no load-bearing self-citation or imported uniqueness theorem is invoked. The registration step relies on a standard external algorithm, and the diffusion stage learns an inverse mapping without the result being tautological to its conditioning inputs by construction. This matches the default expectation for a non-circular paper.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The approach rests on standard assumptions from image registration and generative modeling literature; no new physical entities are introduced and no free parameters are explicitly fitted to the target result beyond typical diffusion training hyperparameters.

axioms (2)
  • domain assumption Phase correlation registration can sufficiently stabilize speckle sequences so that remaining fluctuations reflect flow dynamics
    Invoked as the first step to generate the physics prior before diffusion.
  • domain assumption Conditional diffusion models can invert the imaging process when supplied with a suitable physics prior and limited observations
    Core modeling assumption of the RetinaDiff reconstruction step.

pith-pipeline@v0.9.0 · 5580 in / 1368 out tokens · 66355 ms · 2026-05-10T00:14:28.748222+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

105 extracted references · 105 canonical work pages

  1. [1]

    2008 , publisher=

    Dissipative solitons: from optics to biology and medicine , author=. 2008 , publisher=

    work page 2008

  2. [2]

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , volume=

    Dissipative structures in biological systems: bistability, oscillations, spatial patterns and waves , author=. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , volume=. 2018 , publisher=

    work page 2018

  3. [3]

    2013 , publisher=

    Dissipative solitons in reaction diffusion systems , author=. 2013 , publisher=

    work page 2013

  4. [4]

    2000 , publisher=

    Bose-Einstein condensates and atom lasers , author=. 2000 , publisher=

    work page 2000

  5. [5]

    Nature Reviews Physics , volume=

    Non-equilibrium Bose--Einstein condensation in photonic systems , author=. Nature Reviews Physics , volume=. 2022 , publisher=

    work page 2022

  6. [6]

    Physical Review E—Statistical, Nonlinear, and Soft Matter Physics , volume=

    Creeping solitons in dissipative systems and their bifurcations , author=. Physical Review E—Statistical, Nonlinear, and Soft Matter Physics , volume=. 2007 , publisher=

    work page 2007

  7. [7]

    Laser & Photonics Reviews , volume=

    Spectral pulsations of dissipative solitons in ultrafast fiber lasers: period doubling and beyond , author=. Laser & Photonics Reviews , volume=. 2023 , publisher=

    work page 2023

  8. [8]

    Nature Communications , volume=

    Spontaneous symmetry breaking of dissipative optical solitons in a two-component Kerr resonator , author=. Nature Communications , volume=. 2021 , publisher=

    work page 2021

  9. [9]

    Symmetry broken vectorial Kerr frequency combs from Fabry-P

    Hill, Lewis and Hirmer, Eva-Maria and Campbell, Graeme and Bi, Toby and Ghosh, Alekhya and Del’Haye, Pascal and Oppo, Gian-Luca , journal=. Symmetry broken vectorial Kerr frequency combs from Fabry-P. 2024 , publisher=

    work page 2024

  10. [10]

    Nature Photonics , volume=

    Dissipative solitons for mode-locked lasers , author=. Nature Photonics , volume=. 2012 , publisher=

    work page 2012

  11. [11]

    Optics Express , volume=

    All-normal-dispersion femtosecond fiber laser , author=. Optics Express , volume=. 2006 , publisher=

    work page 2006

  12. [12]

    Laser & Photonics Reviews , volume=

    High-energy femtosecond fiber lasers based on pulse propagation at normal dispersion , author=. Laser & Photonics Reviews , volume=. 2008 , publisher=

    work page 2008

  13. [13]

    Nature , volume=

    Ablation-cooled material removal with ultrafast bursts of pulses , author=. Nature , volume=. 2016 , publisher=

    work page 2016

  14. [14]

    Light: Science & Applications , volume=

    High-contrast, fast chemical imaging by coherent Raman scattering using a self-synchronized two-colour fibre laser , author=. Light: Science & Applications , volume=. 2020 , publisher=

    work page 2020

  15. [15]

    Nature Photonics , volume=

    Frequency comb spectroscopy , author=. Nature Photonics , volume=. 2019 , publisher=

    work page 2019

  16. [16]

    Advances in Optics and Photonics , volume=

    Ultralow-noise mode-locked fiber lasers and frequency combs: principles, status, and applications , author=. Advances in Optics and Photonics , volume=. 2016 , publisher=

    work page 2016

  17. [17]

    Communications Physics , volume=

    Real-time observation of dissipative soliton formation in nonlinear polarization rotation mode-locked fibre lasers , author=. Communications Physics , volume=. 2018 , publisher=

    work page 2018

  18. [18]

    Optics Express , volume=

    Unveiling multi-scale laser dynamics through time-stretch and time-lens spectroscopies , author=. Optics Express , volume=. 2017 , publisher=

    work page 2017

  19. [19]

    Nature Photonics , volume=

    Real-time full-field characterization of transient dissipative soliton dynamics in a mode-locked laser , author=. Nature Photonics , volume=. 2018 , publisher=

    work page 2018

  20. [20]

    Advanced Photonics , volume=

    Revealing the behavior of soliton buildup in a mode-locked laser , author=. Advanced Photonics , volume=. 2019 , publisher=

    work page 2019

  21. [21]

    Optica , volume=

    Observation of soliton explosions in a passively mode-locked fiber laser , author=. Optica , volume=. 2015 , publisher=

    work page 2015

  22. [22]

    Science Advances , volume=

    Breathing dissipative solitons in mode-locked fiber lasers , author=. Science Advances , volume=. 2019 , publisher=

    work page 2019

  23. [23]

    Physical Review Letters , volume=

    Subharmonic entrainment breather solitons in ultrafast lasers , author=. Physical Review Letters , volume=. 2020 , publisher=

    work page 2020

  24. [24]

    invisible

    Visualizing the “invisible” soliton pulsation in an ultrafast laser , author=. Laser & Photonics Reviews , volume=. 2020 , publisher=

    work page 2020

  25. [25]

    Physical Review Letters , volume=

    Revealing the transition dynamics from Q switching to mode locking in a soliton laser , author=. Physical Review Letters , volume=. 2019 , publisher=

    work page 2019

  26. [26]

    Optica , volume=

    Coherence memory and amnesia in a mode-locked Mamyshev oscillator , author=. Optica , volume=. 2024 , publisher=

    work page 2024

  27. [27]

    APL Photonics , volume=

    Phase-encoding of loosely bound soliton molecules , author=. APL Photonics , volume=. 2024 , publisher=

    work page 2024

  28. [28]

    Journal of the Optical Society of America B , volume=

    Relative phase locking of pulses in a passively mode-locked fiber laser , author=. Journal of the Optical Society of America B , volume=. 2003 , publisher=

    work page 2003

  29. [29]

    Physical Review A—Atomic, Molecular, and Optical Physics , volume=

    Discrete family of dissipative soliton pairs in mode-locked fiber lasers , author=. Physical Review A—Atomic, Molecular, and Optical Physics , volume=. 2009 , publisher=

    work page 2009

  30. [30]

    Nature Photonics , volume=

    Ultraweak long-range interactions of solitons observed over astronomical distances , author=. Nature Photonics , volume=. 2013 , publisher=

    work page 2013

  31. [31]

    Nature Photonics , volume=

    All-optical bit storage in a fibre laser by optomechanically bound states of solitons , author=. Nature Photonics , volume=. 2016 , publisher=

    work page 2016

  32. [32]

    Nature Communications , volume=

    Formation of optical supramolecular structures in a fibre laser by tailoring long-range soliton interactions , author=. Nature Communications , volume=. 2019 , publisher=

    work page 2019

  33. [33]

    Optica , volume=

    Noise-mediated Casimir-like pulse interaction mechanism in lasers , author=. Optica , volume=. 2016 , publisher=

    work page 2016

  34. [34]

    Physical Review A—Atomic, Molecular, and Optical Physics , volume=

    Analysis of soliton pattern formation in passively mode-locked fiber lasers , author=. Physical Review A—Atomic, Molecular, and Optical Physics , volume=. 2008 , publisher=

    work page 2008

  35. [35]

    Light: Science & Applications , volume=

    Synthesis and dissociation of soliton molecules in parallel optical-soliton reactors , author=. Light: Science & Applications , volume=. 2021 , publisher=

    work page 2021

  36. [36]

    IEEE Journal of Quantum Electronics , volume=

    Stabilized pulse spacing in soliton lasers due to gain depletion and recovery , author=. IEEE Journal of Quantum Electronics , volume=. 1998 , publisher=

    work page 1998

  37. [37]

    Optics Letters , volume=

    Impact of slow gain dynamics on soliton molecules<? A3B2 show [pmg: line-break justify=" yes"/]?> in mode-locked fiber lasers , author=. Optics Letters , volume=. 2012 , publisher=

    work page 2012

  38. [38]

    Invisible

    Intermittent “Invisible” Soliton Pulsations Induced by the Temporal-Gain Dynamics in Passively Mode-Locked Fiber Lasers , author=. ACS Photonics , volume=. 2024 , publisher=

    work page 2024

  39. [39]

    Light: Science & Applications , volume=

    Dynamic gain driven mode-locking in GHz fiber laser , author=. Light: Science & Applications , volume=. 2024 , publisher=

    work page 2024

  40. [40]

    Light: Science & Applications , volume=

    Understanding low-threshold mode-locking at multi-GHz repetition rate , author=. Light: Science & Applications , volume=. 2025 , publisher=

    work page 2025

  41. [41]

    Scientific Reports , volume=

    Parametric spectro-temporal analyzer (PASTA) for real-time optical spectrum observation , author=. Scientific Reports , volume=. 2013 , publisher=

    work page 2013

  42. [42]

    Physical Review Letters , volume=

    Real-time observation of the buildup of soliton molecules , author=. Physical Review Letters , volume=. 2018 , publisher=

    work page 2018

  43. [43]

    Optics Express , volume=

    Time-stretch probing of ultra-fast soliton dynamics related to Q-switched instabilities in mode-locked fiber laser , author=. Optics Express , volume=. 2018 , publisher=

    work page 2018

  44. [44]

    Physical Review Letters , volume=

    Real-time observation of internal motion within ultrafast dissipative optical soliton molecules , author=. Physical Review Letters , volume=. 2017 , publisher=

    work page 2017

  45. [45]

    Science , volume=

    Real-time spectral interferometry probes the internal dynamics of femtosecond soliton molecules , author=. Science , volume=. 2017 , publisher=

    work page 2017

  46. [46]

    Communications Physics , volume=

    Reconfigurable dynamics of optical soliton molecular complexes in an ultrafast thulium fiber laser , author=. Communications Physics , volume=. 2022 , publisher=

    work page 2022

  47. [47]

    Photonics Research , volume=

    Ultrafast temporal-spectral analysis probes isomeric dynamics in a dissipative soliton resonator , author=. Photonics Research , volume=. 2024 , publisher=

    work page 2024

  48. [48]

    Nature Photonics , volume=

    Resolving the build-up of femtosecond mode-locking with single-shot spectroscopy at 90 MHz frame rate , author=. Nature Photonics , volume=. 2016 , publisher=

    work page 2016

  49. [49]

    Nature Photonics , volume=

    Dispersive Fourier transformation for fast continuous single-shot measurements , author=. Nature Photonics , volume=. 2013 , publisher=

    work page 2013

  50. [50]

    Advances in Physics: X , volume=

    Recent advances on time-stretch dispersive Fourier transform and its applications , author=. Advances in Physics: X , volume=. 2022 , publisher=

    work page 2022

  51. [51]

    Advances in Optics and Photonics , volume=

    Application of space--time duality to ultrahigh-speed optical signal processing , author=. Advances in Optics and Photonics , volume=. 2013 , publisher=

    work page 2013

  52. [52]

    Nature Photonics , volume=

    Single-shot measurement of phase and amplitude by using a heterodyne time-lens system and ultrafast digital time-holography , author=. Nature Photonics , volume=. 2018 , publisher=

    work page 2018

  53. [53]

    Nature Communications , volume=

    Imaging soliton dynamics in optical microcavities , author=. Nature Communications , volume=. 2018 , publisher=

    work page 2018

  54. [54]

    Light: Science & Applications , volume=

    Spatio-temporal breather dynamics in microcomb soliton crystals , author=. Light: Science & Applications , volume=. 2024 , publisher=

    work page 2024

  55. [55]

    Communications Physics , volume=

    Polarization-diverse soliton transitions and deterministic switching dynamics in strongly-coupled and self-stabilized microresonator frequency combs , author=. Communications Physics , volume=. 2024 , publisher=

    work page 2024

  56. [56]

    Communications Physics , volume=

    Nonlocal bonding of a soliton and a blue-detuned state in a microcomb laser , author=. Communications Physics , volume=. 2023 , publisher=

    work page 2023

  57. [57]

    Light: Science & Applications , volume=

    Dissipative soliton generation and real-time dynamics in microresonator-filtered fiber lasers , author=. Light: Science & Applications , volume=. 2022 , publisher=

    work page 2022

  58. [58]

    Nature , volume=

    Recent developments in compact ultrafast lasers , author=. Nature , volume=. 2003 , publisher=

    work page 2003

  59. [59]

    Optics Letters , volume=

    Pulse-forming dynamics of a cw passively mode-locked Ti: sapphire/DDI laser , author=. Optics Letters , volume=. 1992 , publisher=

    work page 1992

  60. [60]

    Photonics Research , volume=

    Self-synchronized temporal-spectral characterization system for revealing ultrafast fiber laser dynamics , author=. Photonics Research , volume=. 2023 , publisher=

    work page 2023

  61. [61]

    Nature Communications , volume=

    Panoramic-reconstruction temporal imaging for seamless measurements of slowly-evolved femtosecond pulse dynamics , author=. Nature Communications , volume=. 2017 , publisher=

    work page 2017

  62. [62]

    Principles of parametric temporal imaging. I. System configurations , author=. IEEE Journal of Quantum Electronics , volume=. 2000 , publisher=

    work page 2000

  63. [63]

    Principles of parametric temporal imaging. II. System performance , author=. IEEE Journal of Quantum Electronics , volume=. 2000 , publisher=

    work page 2000

  64. [64]

    Applied Physics B , volume=

    Passive mode locking with slow saturable absorbers , author=. Applied Physics B , volume=. 2001 , publisher=

    work page 2001

  65. [65]

    Journal of the Optical Society of America B , volume=

    Intermittent events due to spectral filtering induced multi-pulsing instability in a mode-locked fiber laser , author=. Journal of the Optical Society of America B , volume=. 2020 , publisher=

    work page 2020

  66. [66]

    Optics Express , volume=

    From breather soliton molecules to chaos in a laser cavity: the scenario of intermittent transitions , author=. Optics Express , volume=. 2024 , publisher=

    work page 2024

  67. [67]

    Chinese Optics Letters , volume=

    Coherent dissipative soliton intermittency in ultrafast fiber lasers , author=. Chinese Optics Letters , volume=. 2022 , publisher=

    work page 2022

  68. [68]

    Acta Ophthalmologica , volume=

    Retinal microcirculatory parameters-Clinical applications beyond ophthalmology , author=. Acta Ophthalmologica , volume=. 2024 , publisher=

    work page 2024

  69. [69]

    Experimental Eye Research , volume=

    Retinal microcirculation: a window into systemic circulation and metabolic disease , author=. Experimental Eye Research , volume=. 2024 , publisher=

    work page 2024

  70. [70]

    Progress in Retinal and Eye Research , volume=

    Studies of the retinal microcirculation using human donor eyes and high-resolution clinical imaging: insights gained to guide future research in diabetic retinopathy , author=. Progress in Retinal and Eye Research , volume=. 2023 , publisher=

    work page 2023

  71. [71]

    Optics Express , volume=

    Dual-exposure temporal laser speckle imaging for simultaneously accessing microvascular blood perfusion and angiography , author=. Optics Express , volume=. 2024 , publisher=

    work page 2024

  72. [72]

    Optics Communications , volume=

    Laser speckle contrast imaging based on scattering bias factor correction , author=. Optics Communications , volume=. 2024 , publisher=

    work page 2024

  73. [73]

    Tenth Symposium on Novel Optoelectronic Detection Technology and Applications , volume=

    Research progress of laser speckle contrast imaging in dynamic blood flow detection , author=. Tenth Symposium on Novel Optoelectronic Detection Technology and Applications , volume=. 2025 , organization=

    work page 2025

  74. [74]

    Biomedical Optics Express , volume=

    Choosing a model for laser speckle contrast imaging , author=. Biomedical Optics Express , volume=. 2021 , publisher=

    work page 2021

  75. [75]

    Scientific reports , volume=

    Optimizing the precision of laser speckle contrast imaging , author=. Scientific reports , volume=. 2023 , publisher=

    work page 2023

  76. [76]

    Progress in retinal and eye research , volume=

    Quantitative approaches in multimodal fundus imaging: state of the art and future perspectives , author=. Progress in retinal and eye research , volume=. 2023 , publisher=

    work page 2023

  77. [77]

    Journal of Biophotonics , volume=

    Motion artifact correction for OCT microvascular images based on image feature matching , author=. Journal of Biophotonics , volume=. 2024 , publisher=

    work page 2024

  78. [78]

    IEEE Transactions on Image Processing , volume=

    Identification of translational displacements between N-dimensional data sets using the high-order SVD and phase correlation , author=. IEEE Transactions on Image Processing , volume=. 2005 , publisher=

    work page 2005

  79. [79]

    IEEE transactions on information technology in biomedicine , volume=

    Hybrid retinal image registration , author=. IEEE transactions on information technology in biomedicine , volume=. 2006 , publisher=

    work page 2006

  80. [80]

    IEEE transactions on medical imaging , volume=

    Manhattan distance-based adaptive 3D transform-domain collaborative filtering for laser speckle imaging of blood flow , author=. IEEE transactions on medical imaging , volume=. 2019 , publisher=

    work page 2019

Showing first 80 references.