Pith. sign in

REVIEW 3 major objections 7 minor 30 references

Heart Ultrasound Segmentation Hits 93.87% Dice via Transition Probabilities

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · glm-5.2

2026-07-09 06:07 UTC pith:F46COQOJ

load-bearing objection Legitimate engineering contribution to echo segmentation, but SOTA claim is not robustly supported without significance testing. the 3 major comments →

arxiv 2607.07580 v1 pith:F46COQOJ submitted 2026-07-08 cs.CV

Automatic Echocardiography Segmentation via Transition Probability Correlation for Stable Semantic Extraction

Xinran Chen , Xiyuan Wang , Guangquan Zhou , Chuan Chen This is my paper
classification cs.CV
keywords echocardiography segmentationtransition probabilityspatiotemporal featuressemi-supervised learningfrequency-domain pre-trainingspeckle noisecardiac motionlocal self-similarity
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper tries to prove that deep learning segmentation of echocardiography videos can be substantially improved by exploiting a specific kind of local motion cue: inter-frame affinity matrices interpreted as pixel-level transition probability distributions of cardiac motion. The authors argue that these transition probabilities are more stable under speckle noise than raw texture features, and that using them to guide semantic correction produces smoother, more anatomically consistent segmentations. They build a semi-supervised network called CISR-Net around a Spatio-Temporal Local Self-Similarity Fusion (STLSF) module, which first rectifies deep semantic features using local transition probability correlations and then uses those refined semantics to clean up noisy shallow texture features. A frequency-domain denoising pre-training strategy adapts the encoder to ultrasound-specific noise patterns before fine-tuning. On two public benchmarks, CAMUS and EchoNet-Dynamic, the method achieves Dice scores of 93.87% and 92.62% with boundary errors (HD95) of 3.29mm and 2.73mm, surpassing eight competing methods including both general video models and echocardiography-specific networks. The paper's central claim is that transition probability correlations, computed via local linear projections and softmax normalization within spatiotemporal windows, capture structural motion patterns that resist imaging artifacts better than direct feature matching, and that this stability propagates through the network to yield better boundaries.

Core claim

The central mechanism is the Transition-to-Semantics Guidance (TSG) component of the STLSF module. For each pixel in a deep feature map, TSG computes a local transition probability distribution by projecting the pixel and its neighbors through separate linear layers, taking their dot product, and applying softmax. This distribution is treated as an implicit representation of cardiac motion. An attention operation then uses these transition probability distributions as queries and keys while the original features serve as values, so that semantic alignment is driven by structural motion coherence rather than by noise-corrupted texture. The paper demonstrates that this two-step process—compute

What carries the argument

Spatio-Temporal Local Self-Similarity Fusion (STLSF) module with two components: (1) Transition-to-Semantics Guidance (TSG), which computes local inter-frame transition probability distributions via linear projections and softmax, then uses them as structural queries/keys in an attention operation to rectify deep semantic features; (2) Semantics-to-Texture Guidance (STG), which uses the refined deep semantics to guide local cross-attention over shallow texture features for boundary-consistent output. A Frequency-aware Denoising (FD) pre-training strategy injects speckle-like noise in the DCT frequency domain and trains the encoder to recover clean frequency spectra.

Load-bearing premise

The paper assumes that inter-frame affinity matrices computed via local linear projections and softmax meaningfully represent transition probability distributions of cardiac motion and are more stable than raw features under speckle noise. This is stated conceptually but never empirically validated—no analysis shows the computed transition probabilities correlate with actual cardiac motion or are demonstrably more stable than alternatives.

What would settle it

If one replaced the transition probability computation (Eq. 6-7) with a standard local self-attention using raw features as queries, keys, and values—without any motion-probability interpretation—and achieved equivalent segmentation performance, the central claim that transition probability correlations specifically drive the improvement would be undermined.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If transition probability correlations are genuinely more stable than raw features under noise, the same mechanism could improve segmentation in other imaging modalities with heavy speckle or artifact corruption, such as optical coherence tomography or ultrasound of other organs.
  • The frequency-domain pre-training strategy could be extended to incorporate modality-specific spectral priors beyond speckle noise, potentially improving encoder initialization for any imaging system with known noise spectral characteristics.
  • The local-window restriction for transition probability computation keeps computational cost manageable, which could make the approach viable for real-time clinical deployment where 3D or global temporal models are too expensive.
  • The semi-supervised design—temporal constraints from intermediate frames with labels only at end-diastolic and end-systolic frames—suggests a practical path for clinical settings where exhaustive frame annotation is infeasible.

Where Pith is reading between the lines

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

  • The paper does not empirically validate that the computed transition probabilities actually correspond to physical cardiac motion or are more stable than alternatives under noise. If they are simply re-encoded attention affinities relabeled as 'transition probabilities,' the improvement may come from the local windowing and dual-projection design rather than from any motion-modeling property, in w
  • The ablation shows TSG and STG each contribute roughly 0.25% Dice improvement and combine for 0.50%, suggesting the gains are additive but modest. The frequency-domain pre-training contributes a larger share of the total improvement (roughly 0.3-0.4% Dice over alternatives), raising the question of whether the STLSF module's complexity is justified relative to simply investing more in pre-training
  • The method processes 10-frame clips at 150 frames/s, which is clinically practical, but the paper does not report inference latency on variable-length or streaming inputs, leaving open whether the local-window approach scales to continuous acquisition scenarios.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper proposes CISR-Net, a semi-supervised echocardiography segmentation framework with two main contributions: (1) a Spatio-temporal Local Self-Similarity Fusion (STLSF) module that uses local transition probability correlations to rectify noise-induced semantic errors in deep features (TSG) and then guides shallow texture refinement (STG), and (2) a Frequency-aware Denoising (FD) pre-training strategy that injects frequency-domain perturbations to adapt the encoder to ultrasound-specific noise patterns. The backbone is a convolutional network inspired by OverLoCK with a base-overview-focus architecture. The method is evaluated on CAMUS and EchoNet-Dynamic, reporting Dice scores of 93.87% and 92.62% respectively, with ablation studies isolating the contributions of TSG, STG, and FD pre-training.

Significance. The paper addresses a clinically relevant problem and the STLSF module is a reasonable architectural contribution. The frequency-aware denoising pre-training is a sensible domain-specific adaptation. The evaluation covers two standard benchmarks with eight baselines including recent methods (NCMNet, EchoONE, MemSAM), and boundary metrics (HD95, ASSD) are reported alongside Dice. The computational profile (78.62 GFLOPs, 38.5 MB) is practical. However, the magnitude of improvement over the strongest baseline is marginal, and the experimental protocol does not establish that these gains are statistically robust.

major comments (3)
  1. §3.2 and Table 1: The SOTA claim rests on improvements of 0.31% Dice over NCMNet on CAMUS (93.87 vs 93.56) and 0.32% on EchoNet-Dynamic (92.62 vs 92.30). No variance, confidence intervals, or significance tests are reported, and §3.2 states the model is selected based on 'peak validation Dice score,' which introduces optimistic selection bias. Without multi-seed runs, sub-0.5% differences cannot be distinguished from run-to-run variance. This is load-bearing because the paper's central claim is SOTA performance. At minimum, the authors should report results over multiple seeds and provide standard deviations or confidence intervals.
  2. Table 2(a) vs Table 2(b): The 'Ours' row in Table 2(a) reports Dice = 93.25%, while Table 2(b) reports 'Ours' Dice = 93.19%. Both are described as averaged over CAMUS and EchoNet-Dynamic, and both should represent the full model (STLSF + FD pre-training). The 0.06% discrepancy is unexplained and raises questions about whether the two ablation studies used different runs, different evaluation protocols, or different model configurations. The authors should clarify whether these are the same model and, if not, explain the difference.
  3. §2.3, Eq. (6)–(7): The paper frames the inter-frame affinity matrix computed via local linear projections (Eq. 6) as a 'transition probability distribution' that 'implicitly represents cardiac motion,' and claims this representation is more stable than raw features under speckle noise. This is the conceptual basis for the TSG module, but no empirical evidence is provided that the computed affinities actually correlate with physical cardiac motion or are more stable than alternatives (e.g., raw feature similarity, optical flow). The ablation in Table 2(a) shows TSG contributes +0.26% Dice, which demonstrates practical utility, but does not validate the transition-probability interpretation. A simple analysis (e.g., visualizing the affinity matrices or comparing stability under noise against a baseline) would substantially strengthen the central claim.
minor comments (7)
  1. §2.2, Eq. (1): The notation '1/f ultrasound speckle noise map' for N is unclear. Is this a sample from a 1/f power spectral density? The generation procedure for N should be specified.
  2. §2.3, Eq. (7): The attention formulation uses Q(P_trans(m)), K(P_trans(n)), V(F̃_n). It is unclear whether Q, K, V are shared with or distinct from Linear1, Linear2 in Eq. (6). Clarify the parameter relationship.
  3. Table 1: The 'Venue/Year' column lists 'UMB 2024' for SimLVSeg — this appears to be an abbreviation for 'Ultrasound in Medicine & Biology'; consider spelling out the journal name for consistency with other entries.
  4. §3.1: The resolution difference between CAMUS (224×224) and EchoNet-Dynamic (128×128) is noted but its potential impact on cross-dataset comparisons is not discussed. A brief note would help.
  5. Fig. 1: The figure caption is minimal. Labels for sub-figures (a), (b), (c) referenced in the text (e.g., 'Fig. 1(c)') should be explicitly marked in the figure.
  6. §2.3: The local window size k is mentioned in the definition of N(m) but its value is not specified in the experimental setup. This is a free parameter that should be reported.
  7. References [18] (OverLoCK) is cited as the backbone inspiration but the relationship between the proposed architecture and OverLoCK is not detailed enough for reproducibility. Clarify what is borrowed vs. novel.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for a careful and constructive review. The referee raises three major points: (1) the absence of multi-seed runs and significance testing for sub-0.5% Dice improvements, (2) an unexplained 0.06% discrepancy between the 'Ours' rows in Tables 2(a) and 2(b), and (3) the lack of empirical validation for the transition-probability interpretation underlying the TSG module. We agree that all three points require action and will address each in the revised manuscript.

read point-by-point responses

  1. Referee: §3.2 and Table 1: The SOTA claim rests on improvements of 0.31% Dice over NCMNet on CAMUS and 0.32% on EchoNet-Dynamic. No variance, confidence intervals, or significance tests are reported, and the model is selected based on 'peak validation Dice score,' introducing optimistic selection bias. Without multi-seed runs, sub-0.5% differences cannot be distinguished from run-to-run variance.

    Authors: The referee is correct. We will conduct multi-seed experiments (at least 5 seeds with different random initializations) for both CISR-Net and the strongest baselines (NCMNet, EchoONE, MemSAM) on both CAMUS and EchoNet-Dynamic. We will report mean Dice, standard deviation, and paired t-tests (or Wilcoxon signed-rank tests where appropriate). We will also revise the model selection description in §3.2 to clarify that selection is based on validation set performance only, and we will report test-set results corresponding to the best validation epoch. If the significance tests do not support the SOTA claim at conventional thresholds, we will revise the language from 'SOTA performance' to 'competitive performance with improvements in boundary metrics,' since our HD95 and ASSD improvements are more substantial (e.g., 3.29mm vs 4.08mm on CAMUS, 2.73mm vs 4.04mm on EchoNet-Dynamic relative to NCMNet). We acknowledge that the Dice improvements alone are marginal and that the boundary metric improvements, while larger in relative terms, also require statistical validation. revision: yes

  2. Referee: Table 2(a) vs Table 2(b): The 'Ours' row in Table 2(a) reports Dice = 93.25%, while Table 2(b) reports 'Ours' Dice = 93.19%. Both are described as averaged over CAMUS and EchoNet-Dynamic and both should represent the full model. The 0.06% discrepancy is unexplained.

    Authors: The referee has identified a genuine inconsistency. The two ablation tables were conducted at different stages of our experiments: Table 2(a) was run during the STLSF module development phase, while Table 2(b) was run later when evaluating pre-training strategies. Although both use the same model configuration, minor differences in random seeds and the order in which experiments were conducted led to the 0.06% discrepancy. This is not acceptable for a paper claiming precise ablation results. In the revision, we will re-run both ablation studies under identical conditions (same seeds, same data splits, same evaluation protocol) and ensure the full-model rows match exactly. We will also report standard deviations across seeds for all ablation entries, which will contextualize whether 0.06% is within run-to-run variance (which we expect it is). revision: yes

  3. Referee: §2.3, Eq. (6)–(7): The paper frames the inter-frame affinity matrix as a 'transition probability distribution' that 'implicitly represents cardiac motion' and claims this representation is more stable than raw features under speckle noise. No empirical evidence is provided that the computed affinities actually correlate with physical cardiac motion or are more stable than alternatives. The ablation shows TSG contributes +0.26% Dice but does not validate the transition-probability interpretation.

    Authors: This is a fair criticism. The conceptual claim that the affinity matrix represents transition probabilities correlated with cardiac motion is currently asserted but not empirically validated. We will add the following analyses to the revised manuscript: (1) Visualization of the learned affinity matrices P_trans at different cardiac phases (ED, mid-systole, ES, mid-diastole) to show that the attention patterns shift in a manner consistent with known cardiac motion patterns. (2) A quantitative stability comparison: we will add controlled speckle noise at varying levels to the same input frames and measure the variance of the affinity matrices versus raw feature similarity matrices, demonstrating that the transition-probability formulation is more stable under noise perturbation. (3) We will compare TSG against a variant using raw feature similarity (without the linear projection + softmax normalization) to directly test whether the transition-probability formulation provides benefits beyond simple feature matching. If these analyses do not support the stability claim, we will revise the framing to describe TSG as a learned attention-based semantic rectification module without the transition-probability interpretation. revision: yes

Circularity Check

0 steps flagged

No circularity detected; method is evaluated against external benchmarks with standard metrics and ablation studies compare component variants against baselines.

full rationale

The paper's central claims are evaluated against external benchmarks (CAMUS, EchoNet-Dynamic) using standard metrics (Dice, HD95, ASSD). The STLSF module (Eqs. 6-7) is a standard local attention mechanism whose 'transition probability' framing is conceptually inspired by self-citations [25, 21], but the actual computation is a learnable linear-projection-based affinity that is independently trained and evaluated. The citations to [25] (Wang et al., CVPR 2019) and [21] (Shechtman & Irani, CVPR 2007) are external works by different author groups, not self-citations. The FD pre-training (Eqs. 1-2) is a standard denoising autoencoder variant in the frequency domain, independently evaluated in ablation Table 2(b). No step in the derivation chain reduces to its inputs by construction: the transition affinity matrix (Eq. 6) is computed from learned projections of the input features, and the rectified output (Eq. 7) is produced via attention using those affinities as Q/K — this is a standard architectural design, not a definitional identity. The ablation studies (Tables 2a, 2b) compare against baselines without the proposed modules, showing incremental improvements. While the improvements are marginal and lack significance testing (a correctness concern, not a circularity concern), the method is not circular: the outputs are not defined in terms of the targets they claim to predict. The conceptual claim that 'transition probabilities' represent cardiac motion is unvalidated, but this is an empirical assumption, not a circular definition.

Axiom & Free-Parameter Ledger

4 free parameters · 3 axioms · 0 invented entities

The paper introduces no new physical entities or postulated objects. The 'transition probability' is a reinterpretation of attention weights, not a new entity. Free parameters are mostly architectural hyperparameters with unreported values. The axioms are domain assumptions about echocardiography structure and noise that are plausible but unvalidated within the paper.

free parameters (4)
  • λ (noise intensity scale) = not specified
    Controls noise injection magnitude in FD pre-training (Eq. 1). Value not reported.
  • k (local window size) = not specified
    Defines neighborhood N(m) for transition probability computation. Value not reported.
  • Loss weights (0.8, 1.2) = 0.8, 1.2
    L_total = 0.8*L_bce + 1.2*L_dice. Chosen without justification.
  • T (clip length) = 10
    Number of frames sampled per video clip. Chosen without ablation.
axioms (3)
  • domain assumption Inter-frame affinity matrices can be viewed as pixel-level transition probability distributions that capture spatiotemporal evolution
    §2.3, inspired by [25]. The paper assumes the softmax of linear-projected dot products (Eq. 6) represents motion transition probabilities. This is never empirically validated.
  • domain assumption Spatial layout is more stable than texture in echocardiography
    §1, citing Local Self-Similarity [21]. Used to justify the STLSF design but not tested directly.
  • domain assumption Frequency-domain perturbation teaches ultrasound-specific noise patterns
    §2.2. The FD pre-training assumes DCT-spectrum corruption models real speckle noise. The 1/f noise map (Eq. 1) is an approximation of speckle statistics.

pith-pipeline@v1.1.0-glm · 12141 in / 2398 out tokens · 345741 ms · 2026-07-09T06:07:24.233654+00:00 · methodology

0 comments
read the original abstract

While echocardiography is essential for cardiovascular diagnosis, inherent speckle noise and low signal-to-noise ratio often lead to ambiguous semantic features and fragmented boundaries. These limitations significantly hinder the segmentation accuracy of deep learning models in complex clinical cases. Moreover, temporal motion of the heart plays a critical role in recognizing anatomical structures. To address these challenges, we designed a STLSF module which comprises a window-matching-based semantic correction component and a semantics-guided texture enhancement component. By leveraging local transition probability correlations to correct semantics and employing semantics-guided texture enhancement, the STLSF module effectively mitigates texture instability and ambiguous semantic interpretations caused by disadvantaged echocardiography quality. Additionally, to facilitate the encoder's adaptation to the intrinsic priors of ultrasound-specific imaging patterns, we propose a frequency-aware denoising pre-training method. The entire work builds a convolution-based network with locality inductive bias and long-range dependencies. Extensive experiments confirm our SOTA performance, achieving 93.87\% Dice on CAMUS and 92.62\% on EchoNet-Dynamic, with respective HD95 values of 3.29mm and 2.73mm.

Figures

Figures reproduced from arXiv: 2607.07580 by Chuan Chen, Guangquan Zhou, Xinran Chen, Xiyuan Wang.

Figure 1
Figure 1. Figure 1: Overall architecture of the FD pre-training framework and CISR-Net. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Visual comparison with representative baselines on CAMUS and EchoNet [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

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

30 extracted references · 30 canonical work pages

  1. [1]

    IEEE Transactions on Ultrasonics, Ferroelectrics, and Fre- quency Control69(6), 2115–2128 (2022).https://doi.org/10.1109/TUFFC.2022

    Awasthi, N., Vermeer, L., Fixsen, L.S., Lopata, R.G.P., Pluim, J.P.W.: Lvnet: Lightweight model for left ventricle segmentation for short axis views in echocar- diographic imaging. IEEE Transactions on Ultrasonics, Ferroelectrics, and Fre- quency Control69(6), 2115–2128 (2022).https://doi.org/10.1109/TUFFC.2022. 3169684

    work page doi:10.1109/tuffc.2022 2022

  2. [2]

    In: Proceedings of the IEEE/CVF inter- national conference on computer vision

    Bekuzarov, M., Bermudez, A., Lee, J.Y., Li, H.: Xmem++: Production-level video segmentation from few annotated frames. In: Proceedings of the IEEE/CVF inter- national conference on computer vision. pp. 635–644 (2023)

    work page 2023

  3. [3]

    Chen, C., Qin, C., Qiu, H., Tarroni, G., Duan, J., Bai, W., Rueck- ert, D.: Deep learning for cardiac image segmentation: a review. Fron- tiers in Cardiovascular MedicineV olume 7 - 2020(2020).https:// doi.org/10.3389/fcvm.2020.00025,https://www.frontiersin.org/journals/ cardiovascular-medicine/articles/10.3389/fcvm.2020.00025

    work page doi:10.3389/fcvm.2020.00025 2020

  4. [4]

    IEEE Transactions on Medical Imaging42(5), 1289–1300 (2023).https://doi.org/10.1109/TMI.2022

    Chen, G., Li, L., Dai, Y., Zhang, J., Yap, M.H.: AAU-Net: An adaptive attention u-net for breast lesions segmentation in ultrasound images. IEEE Transactions on Medical Imaging42(5), 1289–1300 (2023).https://doi.org/10.1109/TMI.2022. 3226268

    work page doi:10.1109/tmi.2022 2023

  5. [5]

    P., Huang, Z., Romero, A

    Chen,H.,Wang,Y.,Guo,T.,Xu,C.,Deng,Y.,Liu,Z.,Ma,S.,Xu,C.,Xu,C.,Gao, W.: Pre-trained image processing transformer. In: 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 12294–12305 (2021). https://doi.org/10.1109/CVPR46437.2021.01212

    work page doi:10.1109/cvpr46437.2021.01212 2021

  6. [6]

    IEEE Transactions on Medical Imaging44(12), 5172–5182 (2025)

    Deng, X., Wu, H.: Echocardiography video segmentation via neighborhood corre- lation mining. IEEE Transactions on Medical Imaging44(12), 5172–5182 (2025). https://doi.org/10.1109/TMI.2025.3588157

    work page doi:10.1109/tmi.2025.3588157 2025

  7. [7]

    Cross-dimension affinity distillation for 3d em neuron segmentation,

    Deng, X., Wu, H., Zeng, R., Qin, J.: Memsam: Taming segment anything model for echocardiography video segmentation. In: 2024 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR). pp. 9622–9631 (2024).https: //doi.org/10.1109/CVPR52733.2024.00919

    work page doi:10.1109/cvpr52733.2024.00919 2024

  8. [8]

    Circulation: Cardiovascular Imaging17(4), e016517 (2024).https: //doi.org/10.1161/CIRCIMAGING.124.016517,https://www.ahajournals.org/ doi/abs/10.1161/CIRCIMAGING.124.016517

    Gillam, L.D., Marcoff, L.: Echocardiography: Past, present, and fu- ture. Circulation: Cardiovascular Imaging17(4), e016517 (2024).https: //doi.org/10.1161/CIRCIMAGING.124.016517,https://www.ahajournals.org/ doi/abs/10.1161/CIRCIMAGING.124.016517

    work page doi:10.1161/circimaging.124.016517 2024

  9. [9]

    Medical Image Analysis71, 102042 (2021)

    Guo, L., Lei, B., Chen, W., Du, J., Frangi, A.F., Qin, J., Zhao, C., Shi, P., Xia, B., Wang, T.: Dual attention enhancement feature fu- sion network for segmentation and quantitative analysis of paedi- atric echocardiography. Medical Image Analysis71, 102042 (2021). https://doi.org/https://doi.org/10.1016/j.media.2021.102042,https: //www.sciencedirect.com...

    work page doi:10.1016/j.media.2021.102042 2021

  10. [10]

    Computer Meth- ods and Programs in Biomedicine108(3), 1261–1276 (2012).https://doi.org/ https://doi.org/10.1016/j.cmpb.2012.08.011,https://www.sciencedirect

    Hassan,M.,Chaudhry,A.,Khan,A.,Kim,J.Y.:Carotidarteryimagesegmentation using modified spatial fuzzy c-means and ensemble clustering. Computer Meth- ods and Programs in Biomedicine108(3), 1261–1276 (2012).https://doi.org/ https://doi.org/10.1016/j.cmpb.2012.08.011,https://www.sciencedirect. com/science/article/pii/S0169260712001939 10 X. Chen et al

    work page doi:10.1016/j.cmpb.2012.08.011 2012

  11. [11]

    IEEE Transactions on Medical Imaging42(9), 2763–2775 (2023).https://doi.org/10.1109/TMI.2023.3264513

    He, A., Wang, K., Li, T., Du, C., Xia, S., Fu, H.: H2former: An efficient hierarchical hybrid transformer for medical image segmentation. IEEE Transactions on Medical Imaging42(9), 2763–2775 (2023).https://doi.org/10.1109/TMI.2023.3264513

    work page doi:10.1109/tmi.2023.3264513 2023

  12. [12]

    In: IEEE Conf

    Hu, J., Xue, W., Cheng, J., Liu, Y., Zhuo, W., Ni, D.: Echoone: Segmenting multiple echocardiography planes in one model. In: 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 5207–5216 (2025). https://doi.org/10.1109/CVPR52734.2025.00491

    work page doi:10.1109/cvpr52734.2025.00491 2025

  13. [13]

    Journal of the American Soci- ety of Echocardiography28(1), 1–39.e14 (2015).https://doi.org/https://doi

    Lang, R.M., Badano, L.P., Mor-Avi, V., Afilalo, J., Armstrong, A., Ernande, L., Flachskampf, F.A., Foster, E., Goldstein, S.A., Kuznetsova, T., Lancellotti, P., Muraru, D., Picard, M.H., Rietzschel, E.R., Rudski, L., Spencer, K.T., Tsang, W., Voigt, J.U.: Recommendations for cardiac chamber quantification by echocardio- graphy in adults: An update from th...

    work page doi:10.1016/j.echo.2014.10.003 2015

  14. [14]

    IEEE Transactions on Medical Imaging 38(9), 2198–2210 (2019).https://doi.org/10.1109/TMI.2019.2900516

    Leclerc, S., Smistad, E., Pedrosa, J., Østvik, A., Cervenansky, F., Espinosa, F., Espeland, T., Berg, E.A.R., Jodoin, P.M., Grenier, T., Lartizien, C., D’hooge, J., Lovstakken, L., Bernard, O.: Deep learning for segmentation using an open large-scale dataset in 2D echocardiography. IEEE Transactions on Medical Imaging 38(9), 2198–2210 (2019).https://doi.o...

    work page doi:10.1109/tmi.2019.2900516 2019

  15. [15]

    In: Leonardis, A., Ricci, E., Roth, S., Russakovsky, O., Sattler, T., Varol, G

    Li, K., Li, X., Wang, Y., He, Y., Wang, Y., Wang, L., Qiao, Y.: Videomamba: State space model for efficient video understanding. In: Leonardis, A., Ricci, E., Roth, S., Russakovsky, O., Sattler, T., Varol, G. (eds.) Computer Vision – ECCV

    work page

  16. [16]

    pp. 237–255. Springer Nature Switzerland, Cham (2025)

    work page 2025

  17. [17]

    IEEE Transactions on Medical Imaging44(5), 2182–2193 (2025).https: //doi.org/10.1109/TMI.2025.3526955

    Li, X., Cui, C., Shi, S., Fei, H., Hu, Y.: Semi-supervised echocardiography video segmentation via adaptive spatio-temporal tensor semantic awareness and memory flow. IEEE Transactions on Medical Imaging44(5), 2182–2193 (2025).https: //doi.org/10.1109/TMI.2025.3526955

    work page doi:10.1109/tmi.2025.3526955 2025

  18. [18]

    Medical Image Analysis67, 101873 (2021).https://doi.org/https://doi.org/10.1016/j.media.2020.101873, https://www.sciencedirect.com/science/article/pii/S1361841520302371

    Liu, F., Wang, K., Liu, D., Yang, X., Tian, J.: Deep pyramid lo- cal attention neural network for cardiac structure segmentation in two-dimensional echocardiography. Medical Image Analysis67, 101873 (2021).https://doi.org/https://doi.org/10.1016/j.media.2020.101873, https://www.sciencedirect.com/science/article/pii/S1361841520302371

    work page doi:10.1016/j.media.2020.101873 2021

  19. [19]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Lou, M., Yu, Y.: Overlock: An overview-first-look-closely-next convnet with context-mixing dynamic kernels. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 128–138 (2025)

    work page 2025

  20. [20]

    Ultrasound in Medicine & Biology50(12), 1945–1954 (2024)

    Maani, F., Ukaye, A., Saadi, N., Saeed, N., Yaqub, M.: SimLVSeg: Simplifying left ventricular segmentation in 2-d+time echocardiograms with self- and weakly supervised learning. Ultrasound in Medicine & Biology50(12), 1945–1954 (2024). https://doi.org/https://doi.org/10.1016/j.ultrasmedbio.2024.08.023, https://www.sciencedirect.com/science/article/pii/S03...

    work page doi:10.1016/j.ultrasmedbio.2024.08.023 1945

  21. [21]

    Nature 580(7802), 252–256 (2020)

    Ouyang, D., He, B., Ghorbani, A., Yuan, N., Ebinger, J., Langlotz, C., Zou, J., Ashley, E.: Video-based AI for beat-to-beat assessment of cardiac function. Nature 580(7802), 252–256 (2020)

    work page 2020

  22. [22]

    In: 2007 IEEE Conference on Computer Vision and Pattern Recognition

    Shechtman, E., Irani, M.: Matching local self-similarities across images and videos. In: 2007 IEEE Conference on Computer Vision and Pattern Recognition. pp. 1–8 (2007).https://doi.org/10.1109/CVPR.2007.383198

    work page doi:10.1109/cvpr.2007.383198 2007

  23. [23]

    JACC86(22), 2167–2243 (2025).https://doi.org/10.1016/ j.jacc.2025.08.015, focus Issue: Global Burden Of Disease Spotlight

    Stark, B.A., DeCleene, N.K., Desai, E.C., Hsu, J.M., et al.: Global, regional, and national burden of cardiovascular diseases and risk factors in 204 countries and ter- Echocardiography Segmentation via Transition Probability Correlation 11 ritories, 1990-2023. JACC86(22), 2167–2243 (2025).https://doi.org/10.1016/ j.jacc.2025.08.015, focus Issue: Global B...

    work page 1990

  24. [24]

    The Lancet Digital Health4, e46–e54 (2021).https://doi.org/10.1016/S2589-7500(21)00235-1

    Tromp,J.,Seekings,P.,Hung,C.L.,etal.:Automatedinterpretationofsystolicand diastolic function on the echocardiogram: a multicohort study. The Lancet Digital Health4, e46–e54 (2021).https://doi.org/10.1016/S2589-7500(21)00235-1

    work page doi:10.1016/s2589-7500(21)00235-1 2021

  25. [25]

    In: Proceedings of the 25th Interna- tional Conference on Machine Learning

    Vincent, P., Larochelle, H., Bengio, Y., Manzagol, P.A.: Extracting and composing robust features with denoising autoencoders. In: Proceedings of the 25th Interna- tional Conference on Machine Learning. p. 1096–1103. ICML ’08, Association for Computing Machinery, New York, NY, USA (2008).https://doi.org/10.1145/ 1390156.1390294,https://doi.org/10.1145/139...

    work page doi:10.1145/1390156.1390294 2008

  26. [26]

    In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Wang, X., Jabri, A., Efros, A.A.: Learning correspondence from the cycle- consistency of time. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 2561–2571 (2019).https://doi.org/10.1109/ CVPR.2019.00267

    work page arXiv 2019

  27. [27]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Wu, H., Lin, J., Xie, W., Qin, J.: Super-efficient echocardiography video segmen- tation via proxy-and kernel-based semi-supervised learning. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 37, pp. 2803–2811 (2023)

    work page 2023

  28. [28]

    Medical Image Analysis78, 102397 (2022).https://doi.org/https://doi.org/10.1016/j.media.2022.102397, https://www.sciencedirect.com/science/article/pii/S1361841522000494

    Wu, H., Liu, J., Xiao, F., Wen, Z., Cheng, L., Qin, J.: Semi-supervised segmentation of echocardiography videos via noise-resilient spatiotempo- ral semantic calibration and fusion. Medical Image Analysis78, 102397 (2022).https://doi.org/https://doi.org/10.1016/j.media.2022.102397, https://www.sciencedirect.com/science/article/pii/S1361841522000494

    work page doi:10.1016/j.media.2022.102397 2022

  29. [29]

    In: ICLR (2023)

    Xie, J., Li, W., Zhan, X., Liu, Z., Ong, Y.S., Loy, C.C.: Masked frequency modeling for self-supervised visual pre-training. In: ICLR (2023)

    work page 2023

  30. [30]

    IEEE Journal of Biomedical and Health Informatics27(10), 4804– 4815 (2023).https://doi.org/10.1109/JBHI.2023.3293520

    Zhou, G.Q., Zhang, W.B., Shi, Z.Q., Qi, Z.R., Wang, K.N., Song, H., Yao, J., Chen, Y.: Dsanet: Dual-branch shape-aware network for echocardiography segmentation in apical views. IEEE Journal of Biomedical and Health Informatics27(10), 4804– 4815 (2023).https://doi.org/10.1109/JBHI.2023.3293520

    work page doi:10.1109/jbhi.2023.3293520 2023