Pith. sign in

REVIEW 4 major objections 7 minor 1 cited by

Temporal and Spatial Feature Fusion Framework for Dynamic Micro Expression Recognition

T0 review · 4 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read TSFmicro, a dual-stream network that treats the apex-minus-onset frame difference as its temporal signal and fuses it with a shallow positional transformer on the onset frame, reports the best published accuracy on CASME II (87.50% ACC)…

desk verdict Solid empirical architecture paper for micro-expression recognition, but the SOTA claim rests on a test-set-selected fusion choice without error bars—needs independent confirmation before I'd trust the numbers. read the letter →

arxiv 2505.16372 v1 pith:OGFNMQDT submitted 2025-05-22 cs.CV cs.AI

Feng Liu , Bingyu Nan , Xuezhong Qian , Xiaolan Fu This is my paper · ORCID
classification cs.CVcs.AI
keywords micro-expressionrecognitiontemporal-spatialfeaturefusionRetNetdifferenceframemulti-modalCASMEIISAMMCAS(ME)3
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that combining a temporal branch, which extracts motion from a single apex-minus-onset difference image, with a spatial branch that locates facial action via positional embeddings, yields the best published micro-expression recognition results on three standard datasets. The reported scores top all contemporary state-of-the-art methods: 87.50% accuracy on CASME II, 80.88% on SAMM, and 53.29% UF1 with 62.73% UAR on CAS(ME)3. The paper argues that late fusion of the two feature streams, rather than early or sequential fusion, best preserves the complementary 'where-how' information each branch carries.

What carries the argument

The central object is the dual-stream TSFmicro architecture with late element-wise summation of the two feature maps. The temporal stream computes the difference image $f_t = f_{\text{apex}} - f_{\text{onset}}$, embeds it with a Conv Stem, and applies a RetNet block whose retention mechanism $o_n = \sum_{m=1}^{n} \gamma^{n-m}(Q_n e^{in\theta})(K_m e^{im\theta})^\dagger v_m$ imposes an exponential decay over a sequence of tokens, treating the spatial grid as a sequence to model order-dependent dynamics. The spatial stream slices the onset frame into patches, adds learnable position embeddings, and processes them through only two transformer layers to avoid extracting identity information unrelated to motion.

What would settle it

Feed the same architecture a two-channel input of successive frame differences (onset-to-middle and middle-to-apex) and check whether accuracy improves; also randomize the order of the 196 spatial tokens before the RetNet block and see whether performance drops, which would distinguish real temporal modeling from spatial-layout modeling. Re-running all baselines under identical face cropping and augmentation would confirm the reported margins.

Watch

Extended reading notes

Core claim

The paper's central claim is that TSFmicro achieves superior performance over contemporary state-of-the-art methods on CASME II, SAMM, and CAS(ME)3. The temporal branch, fed only the difference between onset and apex frames, contributes most of the recognition accuracy, while the spatial branch adds geographic localization that improves accuracy particularly for under-represented emotion classes; the ablation shows the full model raises 5-class accuracy on CASME II by 24.20 points over the spatial branch alone. Late fusion by element-wise summation in high-dimensional feature space outperforms early and sequential fusion strategies, which the paper attributes to avoiding interference between temporal and spatial information.

Load-bearing premise

All temporal modeling rests on a single image formed by subtracting the onset frame from the apex frame; if that difference does not capture the motion, the temporal branch is just a static classifier on a transformed image.

Editorial extensions

If this is right

  • A two-frame difference (apex minus onset) together with a spatial branch is sufficient to beat full-video and optical-flow methods, suggesting that complete frame sequences may be unnecessary for micro-expression recognition.
  • Late fusion, summing features in a high-dimensional space, is the best of the four fusion modes tested, implying that delaying cross-modal interaction until the final classifier helps this task.
  • The deliberate use of a shallow two-layer spatial branch indicates that deeper spatial backbones are not beneficial and may introduce identity-specific artifacts.
  • The same framework reaches 94.84% ACC on CASME II's 3-class task and 70.75% UF1 on CAS(ME)3's 4-class task, substantially reducing the gap to practical deployment.

Reading between the lines

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

  • Because the temporal input is a single image, the word 'dynamic' overstates the modeling; the gain likely comes from the difference map acting as a motion prior, and a testable extension would replace RetNet with a plain ResNet on the same difference image to compare.
  • The retention mechanism's decay may operate on spatial patch order rather than true time; randomizing the 196 token positions before the RetNet block would reveal whether the ordering is load-bearing.
  • The paper attributes fusion-strategy differences between CASME II and SAMM to culture, but class imbalance and face-alignment difficulty are not controlled; a balanced resampling experiment could separate those factors.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The manuscript proposes TSFmicro, a dual-stream framework for micro-expression recognition. The temporal branch computes a difference image between the apex and onset frames, processes it through a convolutional stem and a RetNet block, and the spatial branch tokenizes the onset frame with position embeddings through a two-layer transformer. The paper evaluates four fusion strategies, selects 'T-S late' as the final model, and reports state-of-the-art accuracy on CASME II (87.50% 5-class ACC), SAMM (80.88% 5-class ACC), and CAS(ME)3 (53.29% UF1, 62.73% UAR for 7-class), together with ablation studies and visualizations. The central claim is that this architecture and fusion method outperform contemporary methods on three widely used micro-expression datasets.

Significance. If the reported results are reliable, TSFmicro would be a competitive approach for micro-expression recognition, and the paper would provide a useful comparison of temporal-spatial fusion strategies. The ablation clearly shows that the difference-image temporal branch carries most of the classification signal, and the fusion study explores a relevant design space. However, the evidence is currently not strong enough to support the state-of-the-art claim: no error bars or significance tests are reported, the final fusion strategy is selected on the same test folds that produce the headline numbers, and on SAMM the margin over the second-best fusion mode is one to two samples. The paper also does not release code or per-fold details, which limits reproducibility and independent verification.

major comments (4)
  1. [Section 4.4, Tables 4, 5, 7] The final model is selected by evaluating four fusion strategies (Table 7) on the same LOSO folds that are then used for the headline comparisons in Tables 4 and 5. Since T-S(late) is the empirical maximum among four options on those folds, the reported SOTA margins are optimistically biased. Concretely, on SAMM 5-class the gap between T-S(late) and S-to-T is 1.47 percentage points (2 samples out of 136), and the gap to MMNet is 0.74 percentage points (1 sample out of 136); on CASME II 5-class the gap to T-to-S is 1.62 points (4 samples out of 248) and to MMNet is 1.41 points (4 samples out of 248). These margins are within the range of random variation for a paired design at these sample sizes. Please report per-fold results, confidence intervals, and a paired significance test (e.g., Wilcoxon signed-rank or McNemar), and use a nested or validation-based selection procedure for the fusion mode.
  2. [Section 3.1, Eqs. (1)-(2)] The temporal branch is defined by Eq. (1) as a single difference image f_t = f_apex - f_onset. After the Conv Stem, this one image is passed through the RetNet-Block, and Eq. (2) is written as a retention sum over sequence positions n and m. With a single input image, the sum collapses to a single term; no frame sequence is processed. The 'temporal' modeling is therefore a two-frame subtraction followed by static feature extraction, not a temporal sequence model. This weakens the paper's claims of capturing 'temporal relations' and 'dynamic micro-expression recognition.' Either feed an actual sequence of frames (e.g., multiple difference images or a short clip) into the retention mechanism, or revise the text to describe the temporal branch as a motion-difference representation.
  3. [Section 4.2, Tables 3-5] The fair comparison with baselines is not established. Section 4.1.2 states that baseline source codes were used and tuned, but only the MMNet replication is described in Section 4.2. It is unclear whether the other baseline numbers in Tables 3-5 were reproduced with the same face cropping, resizing, augmentation, and LOSO splits as TSFmicro, or taken from the original papers. Since the SOTA margins are a few samples on CASME II and SAMM, small preprocessing differences could change the ranking. Please provide the exact evaluation protocol for each baseline, including the split identity (e.g., subject IDs) and preprocessing pipeline, and ideally release the evaluation code.
  4. [Table 6 ablation] The ablation in Table 6 reports single runs without variance estimates. The main fusion gain on CASME II 5-class is +3.63 ACC (from 83.87 to 87.50); on SAMM 5-class it is +5.88 ACC (from 75.00 to 80.88). These gains are presented as definitive, but on 248 and 136 samples respectively they may not be statistically significant. Please add per-fold breakdowns and significance tests for the ablation, or at least standard deviations across LOSO folds.
minor comments (7)
  1. [Section 1] The sentence 'The total data volume of the 12 published spontaneous micro-expressions datasets ... is about 10,000, which is a typical classification and recognition task compared to other small-sample problem compared to other classification recognition tasks' is garbled and should be rewritten.
  2. [Section 3.1] Eq. (2) uses Q_n, K_m, and V_m without defining these quantities; please add a sentence explaining that they are linear projections of the input token sequence, and clarify how the retention mechanism is applied when the sequence length is one.
  3. [Section 4.1.3] The paper states 'a five-layer RetNet network was employed' and 'a ViT was used' but does not specify the number of heads, hidden dimensions, or the patch size P used in the spatial branch; please provide the full architecture hyperparameters.
  4. [Figures 3 and 5] The phrase 'curve-fitted the scores' is unclear; if these are just training curves or smoothed validation curves, please state that explicitly and describe the smoothing method.
  5. [Section 5 Discussion] The cross-cultural explanation for the difference between T-to-S and S-to-T on CASME II vs SAMM is speculative; the paper does not provide per-subject cultural subgroup analysis or statistical evidence. Please temper the claim or add supporting analysis.
  6. [Table 8] The columns for CASME II include 'Repression' while Table 1 uses 'Rep'; please standardize the label names.
  7. [References] Reference [70] is co-authored by one of the present authors (X. Fu) and is cited only for a general statement about cultural differences; consider removing or citing a non-self reference, or explicitly noting the overlap.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: the architecture and ablation results are self-contained, and the only author-overlapping citation is non-load-bearing.

full rationale

TSFmicro's derivation chain is empirical and does not reduce to its own outputs. The temporal input is defined in Eq. 1 as the difference image f_apex - f_onset; Eqs. 2-5 apply a standard retention/convolution stem, Eqs. 6-8 patch the onset frame with position embeddings, and Eq. 9 is an element-wise fusion. None of these equations is defined in terms of the reported ACC/UF1/UAR values, and no fitted parameter is relabeled as a prediction. The only citation overlapping with the present authorship is [70] (Jingting Li et al., co-authored by Xiaolan Fu), which is used in Section 4.2 solely for the qualitative cross-cultural observation about expression and suppression; it supplies no uniqueness theorem, no architectural constraint, and no fitted number, so it is not load-bearing. A related but distinct methodological concern, not a circularity, is that Section 4.4 selects the T-S(late) fusion mode from the same LOSO folds that produce the headline results, so the reported SOTA margin is the maximum of four correlated variants; this threatens generalizability, but it does not make the derivation equivalent to its inputs.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The paper is an empirical architecture paper. Its central claim rests on hand-chosen hyperparameters (RetNet depth, spatial depth, fusion width, patch size) and on several unvalidated domain assumptions: that Apex minus Onset difference frames capture temporal dynamics, that shallow transformers avoid identity leakage, and that standard components (RetNet, LayerNorm, FC layers) transfer to micro-expression recognition. No new entities are introduced. The most consequential hidden assumption is the single-frame 'temporal' representation.

free parameters (5)
  • RetNet depth = 5 layers
    Chosen without ablation; the temporal branch's capacity and reported accuracy depend on it.
  • Spatial branch depth = 2 transformer layers
    Chosen to limit identity information, but no experiment measures identity leakage versus depth.
  • Fusion layer width = 1024
    FC layer width taken from [55] without task-specific tuning; used for all reported results.
  • Patch size P
    Equation (6) defines patch-based tokenization with patch edge length P, but the paper never states P's value, so exact reproduction is impossible.
  • Training schedule = lr=0.0008, batch=32, 50 epochs
    Standard AdamW settings chosen by hand; no sensitivity analysis, and results depend on them.
assumptions (5)
  • ad hoc to paper The difference between Apex and Onset frames (Eq. 1) captures the temporal characteristics of micro-expression movement.
    Section 3.1 states this without evidence; the framework processes a single difference image, so temporal dynamics over the sequence are not actually modeled.
  • ad hoc to paper Using only two Transformer layers prevents the spatial branch from extracting identity information unrelated to micro-expression motion.
    Section 3.2 asserts this; no experiment measures identity leakage versus depth.
  • domain assumption LayerNorm introduces strong nonlinearities that enhance representational capability of the spatial branch.
    Section 3.2 relies on [54] for this; it is background knowledge, not validated in this task.
  • domain assumption Fully connected layers produce broader feature distributions useful for classification.
    Section 3.3 cites [55]; assumed transferable to MER.
  • domain assumption RetNet retention mechanism (Eq. 2) is effective for sequence modeling when applied to visual tokens.
    The temporal branch feeds a single difference-frame token into RetNet; the retention mechanism was designed for language sequences, and its benefit for a single-frame input is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Temporal and Spatial Feature Fusion Framework for Dynamic Micro Expression Recognition." pith.science (2026). https://pith.science/paper/OGFNMQDT

@misc{pith2026250516372,
  author       = {Pith},
  title        = {Pith review of: Temporal and Spatial Feature Fusion Framework for Dynamic Micro Expression Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OGFNMQDT}},
  note         = {Machine review of arXiv:2505.16372}
}
read the original abstract

When emotions are repressed, an individual's true feelings may be revealed through micro-expressions. Consequently, micro-expressions are regarded as a genuine source of insight into an individual's authentic emotions. However, the transient and highly localised nature of micro-expressions poses a significant challenge to their accurate recognition, with the accuracy rate of micro-expression recognition being as low as 50%, even for professionals. In order to address these challenges, it is necessary to explore the field of dynamic micro expression recognition (DMER) using multimodal fusion techniques, with special attention to the diverse fusion of temporal and spatial modal features. In this paper, we propose a novel Temporal and Spatial feature Fusion framework for DMER (TSFmicro). This framework integrates a Retention Network (RetNet) and a transformer-based DMER network, with the objective of efficient micro-expression recognition through the capture and fusion of temporal and spatial relations. Meanwhile, we propose a novel parallel time-space fusion method from the perspective of modal fusion, which fuses spatio-temporal information in high-dimensional feature space, resulting in complementary "where-how" relationships at the semantic level and providing richer semantic information for the model. The experimental results demonstrate the superior performance of the TSFmicro method in comparison to other contemporary state-of-the-art methods. This is evidenced by its effectiveness on three well-recognised micro-expression datasets.

Figures

Figures reproduced from arXiv: 2505.16372 by the authors.

Figure 1
Figure 1. This paper attempts to consider the potential impact of temporal fusion on micro-expression recognition performance from the perspective of modal fusion. (a) Temporal information between frames is extracted as a temporal feature using difference frames. (b) Position embedding is utilized to learn the positional information associated with the occurrence of an action in order to map it to temporal features. (c) Extra… view at source ↗
Figure 2
Figure 2. An overview of the proposed TSFmicro architecture is presented below. (a)The process of TSFmicro is outlined as follows: firstly, the face is cropped; secondly, the difference frames between Apex and Onset frames are used as the temporal information and Onset frames are used as the spatial information; thirdly, the spatio-temporal sub-branch captures and fuses the spatio-temporal information; and finally, the data i… view at source ↗
Figure 3
Figure 3. Evaluation scores of SAMM, CASME II and CAS(ME)3 datasets under 3/4-class and 5/7-class classification conditions during TSFmicro training [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Confusion matrix evaluation results of our proposed TSFmicro framework with different datasets. MMNet result was obtained by replicating the original author’s steps on a Nvidia 4060 Ti GPU. In addition, we curve-fitted the scores during the TSFmi￾cro training process, …
Figure 5
Figure 5. Figure 5: Evaluation scores of different fusion methods on the SAMM and CASME II datasets under 5 classification conditions during TSFmicro training. T-S (early) is a pre-fusion strategy that fuses temporal and spatial information prior to the recognition backbone following a si…
Figure 6
Figure 6. Figure 6: The feature distributions of the three TSFmicro variants are visualized for three different datasets. S can capture the temporal relation of eyebrow gathering, but it cannot focus on the eyebrow region. S to T can correctly map the temporal relation to the correspondin…
Figure 7
Figure 7. Figure 7: The visualization of feature heatmaps corresponds to the different categories in the SAMM dataset [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. AULLM++: Structured-Token-Conditioned Large Language Models for Micro-Expression Action Unit Detection

    cs.CV 2026-03 conditional novelty 5.5 of 10

    AULLM++ fuses multi-granularity visual tokens with FACS-prior AU graph instructions into an LLM prompt and uses counterfactual consistency training to improve micro-expression AU detection and cross-domain Macro-F1.

Reference graph

Works this paper leans on

72 extracted references · 71 canonical work pages · cited by 1 Pith paper

  1. [1]

    Face2exp: Combating data biases for facial expression recognition

    Dan Zeng, Zhiyuan Lin, Xiao Yan, Yuting Liu, Fei Wang, and Bo Tang. Face2exp: Combating data biases for facial expression recognition. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 20259–20268, 2022

    work page 2022

  2. [2]

    Multi-modal sentiment recognition with residual gating network and emotion intensity attention

    Yadi Wang, Xiaoding Guo, Xianhong Hou, Zhijun Miao, Xiaojin Yang, and Jinkai Guo. Multi-modal sentiment recognition with residual gating network and emotion intensity attention. Neural Networks, 188:107483, 2025

    work page 2025

  3. [3]

    A multi- domain constraint learning system inspired by adaptive cognitive graphs for emotion recognition.Neural Networks, 188:107457, 2025

    Dongrui Gao, Mengwen Liu, Haokai Zhang, Manqing Wang, Hongli Chang, Gaoxiang Ouyang, Shihong Liu, and Pengrui Li. A multi- domain constraint learning system inspired by adaptive cognitive graphs for emotion recognition.Neural Networks, 188:107457, 2025

    work page 2025

  4. [4]

    Tackling micro-expression data shortage via dataset alignment and active learning

    Xianye Ben, Chen Gong, Tianhuan Huang, Chuanye Li, Rui Yan, and Yujun Li. Tackling micro-expression data shortage via dataset alignment and active learning. IEEE Transactions on Multimedia, 25:5429–5443, 2023

    work page 2023

  5. [5]

    Multi-task mid-level feature learning for micro-expression recognition

    Jiachi He, Jian-Fang Hu, Xi Lu, and Wei-Shi Zheng. Multi-task mid-level feature learning for micro-expression recognition. Pattern Recognition, 66:44–52, 2017

    work page 2017

  6. [6]

    Blended emotion in-the-wild: Multi- label facial expression recognition using crowdsourced annotations and deep locality feature learning

    Shan Li and Weihong Deng. Blended emotion in-the-wild: Multi- label facial expression recognition using crowdsourced annotations and deep locality feature learning. International Journal of Computer Vision, 127(6):884–906, 2019

    work page 2019

  7. [7]

    Me-plan: A deep prototypical learning with local attention network for dynamic micro- expression recognition

    Sirui Zhao, Huaying Tang, Shifeng Liu, Yangsong Zhang, Hao Wang, Tong Xu, Enhong Chen, and Cuntai Guan. Me-plan: A deep prototypical learning with local attention network for dynamic micro- expression recognition. Neural Networks, 153:427–443, 2022

    work page 2022

  8. [8]

    A unified and interpretable emo- tion representation and expression generation

    Reni Paskaleva, Mykyta Holubakha, Andela Ilic, Saman Motamed, Luc Van Gool, and Danda Paudel. A unified and interpretable emo- tion representation and expression generation. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 2447–2456, 2024

    work page 2024

Show all 72 references
  1. [9]

    Facial micro-expression recognition using stochastic graph convolutional network and dual transferred learning

    Hui Tang and Li Chai. Facial micro-expression recognition using stochastic graph convolutional network and dual transferred learning. Neural Networks, 178:106421, 2024

    2024

  2. [10]

    Rethinking the learning paradigm for dynamic facial expression recognition

    Hanyang Wang, Bo Li, Shuang Wu, Siyuan Shen, Feng Liu, Shouhong Ding, and Aimin Zhou. Rethinking the learning paradigm for dynamic facial expression recognition. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 17958–17968, 2023

    2023

  3. [11]

    Domain knowledge enhanced vision-language pretrained model for dynamic facial expression recognition

    Liupeng Li, Yuhua Zheng, Shupeng Liu, Xiaoyin Xu, and Taihao Li. Domain knowledge enhanced vision-language pretrained model for dynamic facial expression recognition. In Proceedings of the 32nd ACM International Conference on Multimedia, MM ’24, page 5673–5682, New York, NY , ...

    2024

  4. [12]

    Feature refinement: An expression-specific feature learn- ing and fusion method for micro-expression recognition

    Ling Zhou, Qirong Mao, Xiaohua Huang, Feifei Zhang, and Zhihong Zhang. Feature refinement: An expression-specific feature learn- ing and fusion method for micro-expression recognition. Pattern Recognition, 122:108275, 2022. Feng Liu et al.: Preprint submitted to Elsevier Page ...

    2022

  5. [13]

    Joint local and global information learning with single apex frame detection for micro- expression recognition

    Yante Li, Xiaohua Huang, and Guoying Zhao. Joint local and global information learning with single apex frame detection for micro- expression recognition. IEEE Transactions on Image Processing, 30:249–263, 2021

    2021

  6. [14]

    Facial action unit-based deep learning framework for spotting macro- and micro-expressions in long video sequences

    Bo Yang, Jianming Wu, Zhiguang Zhou, Megumi Komiya, Koki Kishimoto, Jianfeng Xu, Keisuke Nonaka, Toshiharu Horiuchi, Satoshi Komorita, Gen Hattori, Sei Naito, and Yasuhiro Takishima. Facial action unit-based deep learning framework for spotting macro- and micro-expressions in ...

    2021

  7. [15]

    Xuri Ge, Junchen Fu, Fuhai Chen, Shan An, Nicu Sebe, and Joe- mon M. Jose. Towards end-to-end explainable facial action unit recognition via vision-language joint learning. In Proceedings of the 32nd ACM International Conference on Multimedia, MM ’24, page 8189–8198, New York,...

    2024

  8. [16]

    Cmnet: contrastive magnification network for micro-expression recognition

    Mengting Wei, Xingxun Jiang, Wenming Zheng, Yuan Zong, Cheng Lu, and Jiateng Liu. Cmnet: contrastive magnification network for micro-expression recognition. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 119–127, 2023

    2023

  9. [17]

    Meta-mmfnet: Meta-learning-based multi-model fusion network for micro-expression recognition

    Wenjuan Gong, Yue Zhang, Wei Wang, Peng Cheng, and Jordi Gonzàlez. Meta-mmfnet: Meta-learning-based multi-model fusion network for micro-expression recognition. ACM Trans. Multimedia Comput. Commun. Appl., 20(2), September 2023

    2023

  10. [18]

    A review of research on micro-expression recognition algorithms based on deep learning

    Fan Zhang and Lin Chai. A review of research on micro-expression recognition algorithms based on deep learning. Neural Computing and Applications, 36(29):17787–17828, 2024

    2024

  11. [19]

    Casme database: A dataset of spontaneous micro- expressions collected from neutralized faces

    Wen-Jing Yan, Qi Wu, Yong-Jin Liu, Su-Jing Wang, and Xi- aolan Fu. Casme database: A dataset of spontaneous micro- expressions collected from neutralized faces. In 2013 10th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG), pages 1–7, 2013

    2013

  12. [20]

    Casme ii: An improved spontaneous micro- expression database and the baseline evaluation.PLOS ONE, 9(1):1– 8, 01 2014

    Wen-Jing Yan, Xiaobai Li, Su-Jing Wang, Guoying Zhao, Yong-Jin Liu, Yu-Hsin Chen, et al. Casme ii: An improved spontaneous micro- expression database and the baseline evaluation.PLOS ONE, 9(1):1– 8, 01 2014

    2014

  13. [21]

    Cas(me) 2 : A database for spontaneous macro- expression and micro-expression spotting and recognition

    Fangbing Qu, Su-Jing Wang, Wen-Jing Yan, He Li, Shuhang Wu, and Xiaolan Fu. Cas(me) 2 : A database for spontaneous macro- expression and micro-expression spotting and recognition. IEEE Transactions on Affective Computing, 9(4):424–436, 2018

    2018

  14. [22]

    Cas(me)3: A third generation facial spontaneous micro-expression database with depth information and high ecological validity

    Jingting Li, Zizhao Dong, Shaoyuan Lu, Su-Jing Wang, Wen-Jing Yan, Yinhuan Ma, et al. Cas(me)3: A third generation facial spontaneous micro-expression database with depth information and high ecological validity. IEEE Transactions on Pattern Analysis and Machine Intelligence, ...

    2023

  15. [23]

    A spontaneous micro-expression database: Inducement, collection and baseline

    Xiaobai Li, Tomas Pfister, Xiaohua Huang, Guoying Zhao, and Matti Pietikäinen. A spontaneous micro-expression database: Inducement, collection and baseline. In 2013 10th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG), pages 1–6, 2013

    2013

  16. [24]

    Micro-expression spotting: A new benchmark

    Thuong-Khanh Tran, Quang-Nhat V o, Xiaopeng Hong, Xiaobai Li, and Guoying Zhao. Micro-expression spotting: A new benchmark. Neurocomputing, 443:356–368, 2021

    2021

  17. [25]

    4dme: A spontaneous 4d micro-expression dataset with multimodalities.IEEE Transactions on Affective Computing, 14(4):3031–3047, 2023

    Xiaobai Li, Shiyang Cheng, Yante Li, Muzammil Behzad, Jie Shen, Stefanos Zafeiriou, Maja Pantic, and Guoying Zhao. 4dme: A spontaneous 4d micro-expression dataset with multimodalities.IEEE Transactions on Affective Computing, 14(4):3031–3047, 2023

    2023

  18. [26]

    Davison, Cliff Lansley, Nicholas Costen, Kevin Tan, and Moi Hoon Yap

    Adrian K. Davison, Cliff Lansley, Nicholas Costen, Kevin Tan, and Moi Hoon Yap. Samm: A spontaneous micro-facial movement dataset. IEEE Transactions on Affective Computing, 9(1):116–129, 2018

    2018

  19. [27]

    Samm long videos: A spontaneous facial micro- and macro-expressions dataset

    Chuin Hong Yap, Connah Kendrick, and Moi Hoon Yap. Samm long videos: A spontaneous facial micro- and macro-expressions dataset. In 2020 15th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2020), pages 771–776, 2020

    2020

  20. [28]

    Video-based facial micro- expression analysis: A survey of datasets, features and algorithms

    Xianye Ben, Yi Ren, Junping Zhang, Su-Jing Wang, Kidiyo Kpalma, Weixiao Meng, and Yong-Jin Liu. Video-based facial micro- expression analysis: A survey of datasets, features and algorithms. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(9):5826–5846, 2022

    2022

  21. [29]

    Dfme: A new benchmark for dynamic facial micro-expression recognition

    Sirui Zhao, Huaying Tang, Xinglong Mao, Shifeng Liu, Yiming Zhang, Hao Wang, Tong Xu, and Enhong Chen. Dfme: A new benchmark for dynamic facial micro-expression recognition. IEEE Transactions on Affective Computing, 15(3):1371–1386, 2024

    2024

  22. [30]

    in the wild

    Petr Husak, Jan Cech, and Jiri Matas. Spotting facial micro- expressions “in the wild”. In Proc. Computer Vision Winter Workshop, 2017. https://cmp.felk.cvut.cz/~cechj/ ME/

    2017

  23. [31]

    Block division convolutional network with implicit deep features augmentation for micro-expression recognition

    Bin Chen, Kun-Hong Liu, Yong Xu, Qing-Qiang Wu, and Jun- Feng Yao. Block division convolutional network with implicit deep features augmentation for micro-expression recognition. IEEE Transactions on Multimedia, 25:1345–1358, 2023

    2023

  24. [32]

    Context transformer with multiscale fusion for robust facial emotion recogni- tion

    Yanling Gan, Luhui Xu, Shuxiang Song, and Xiaomei Tao. Context transformer with multiscale fusion for robust facial emotion recogni- tion. Pattern Recognition, 167:111720, 2025

    2025

  25. [33]

    Visual-to-eeg cross- modal knowledge distillation for continuous emotion recognition

    Su Zhang, Chuangao Tang, and Cuntai Guan. Visual-to-eeg cross- modal knowledge distillation for continuous emotion recognition. Pattern Recognition, 130:108833, 2022

    2022

  26. [34]

    Subrama- nian

    Anima Majumder, Laxmidhar Behera, and Venkatesh K. Subrama- nian. Emotion recognition from geometric facial features using self-organizing map. Pattern Recognition, 47(3):1282–1293, 2014. Handwriting Recognition and other PR Applications

    2014

  27. [35]

    Facial expression and emotion

    Paul Ekman. Facial expression and emotion. American psychologist, 48(4):384, 1993

    1993

  28. [36]

    Automatic micro-expression recognition from long video using a single spotted apex

    Sze-Teng Liong, John See, KokSheik Wong, and Raphael Chung-Wei Phan. Automatic micro-expression recognition from long video using a single spotted apex. In Chu-Song Chen, Jiwen Lu, and Kai-Kuang Ma, editors, Computer Vision – ACCV 2016 Workshops, pages 345– 360, Cham, 2017. Sp...

    2016

  29. [37]

    Knowledge-driven self-supervised representation learning for facial action unit recognition

    Yanan Chang and Shangfei Wang. Knowledge-driven self-supervised representation learning for facial action unit recognition. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 20385–20394, 2022

    2022

  30. [38]

    A spontaneous micro-expression database: Inducement, collection and baseline

    Xiaobai Li, Tomas Pfister, Xiaohua Huang, Guoying Zhao, and Matti Pietikäinen. A spontaneous micro-expression database: Inducement, collection and baseline. In 2013 10th IEEE International Conference and Workshopson Automatic face and gesture recognition (fg), pages 1–6. IEEE, 2013

    2013

  31. [39]

    Phan, and Yee-Hui Oh

    Yandan Wang, John See, Raphael C.-W. Phan, and Yee-Hui Oh. Lbp with six intersection points: Reducing redundant information in lbp-top for micro-expression recognition. In Daniel Cremers, Ian Reid, Hideo Saito, and Ming-Hsuan Yang, editors, Computer Vision – ACCV 2014, pages 5...

    2014

  32. [40]

    Micro- expression recognition using color spaces

    Su-Jing Wang, Wen-Jing Yan, Xiaobai Li, Guoying Zhao, Chun- Guang Zhou, Xiaolan Fu, Minghao Yang, and Jianhua Tao. Micro- expression recognition using color spaces. IEEE Transactions on Image Processing, 24(12):6034–6047, 2015

    2015

  33. [41]

    Merastc: Micro-expression recognition using effective feature encodings and 2d convolutional neural network

    Puneet Gupta. Merastc: Micro-expression recognition using effective feature encodings and 2d convolutional neural network. IEEE Transactions on Affective Computing, 14(2):1431–1441, 2023

    2023

  34. [42]

    Objective class-based micro-expression recognition un- der partial occlusion via region-inspired relation reasoning network

    Qirong Mao, Ling Zhou, Wenming Zheng, Xiuyan Shao, and Xiao- hua Huang. Objective class-based micro-expression recognition un- der partial occlusion via region-inspired relation reasoning network. IEEE Transactions on Affective Computing, 13(4):1998–2016, 2022

    1998

  35. [43]

    A main directional mean optical flow feature for spontaneous micro-expression recognition

    Yong-Jin Liu, Jin-Kai Zhang, Wen-Jing Yan, Su-Jing Wang, Guoying Zhao, and Xiaolan Fu. A main directional mean optical flow feature for spontaneous micro-expression recognition. IEEE Transactions on Affective Computing, 7(4):299–310, 2016

    2016

  36. [44]

    Hybrid facial regions extraction for micro-expression recogni- tion system

    Sze Liong, John See, Raphael Phan, Koksheik Wong, and Su-Wei Tan. Hybrid facial regions extraction for micro-expression recogni- tion system. Journal of Signal Processing Systems, 90:601–617, 04 2018. Feng Liu et al.: Preprint submitted to Elsevier Page 15 of 16 Temporal and S...

    2018

  37. [45]

    Sze-Teng Liong, Y . S. Gan, John See, Huai-Qian Khor, and Yen- Chang Huang. Shallow triple stream three-dimensional cnn (ststnet) for micro-expression recognition. In 2019 14th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2019), pages 1–5, 2019

    2019

  38. [46]

    Y . S. Gan, Sung-En Lien, Yi-Chen Chiang, and Sze-Teng Liong. Laenet for micro-expression recognition. The Visual Computer, 40(2):585–599, March 2023

    2023

  39. [47]

    Sze-Teng Liong, Y . S. Gan, Danna Zheng, Shu-Meng Li, Hao-Xuan Xu, Han-Zhe Zhang, Ran-Ke Lyu, and Kun-Hong Liu. Evaluation of the spatio-temporal features and gan for micro-expression recog- nition system. Journal of Signal Processing Systems, 92(7):705–725, July 2020

    2020

  40. [48]

    Multi- scale active patches fusion based on spatiotemporal lbp-top for micro-expression recognition

    Zhe Sun, Zheng ping Hu, Mengyao Zhao, and Shufang Li. Multi- scale active patches fusion based on spatiotemporal lbp-top for micro-expression recognition. Journal of Visual Communication and Image Representation, 71:102862, 2020

    2020

  41. [49]

    Micro- expression action unit detection with spatial and channel attention

    Yante Li, Xiaohua Huang, and Guoying Zhao. Micro- expression action unit detection with spatial and channel attention. Neurocomputing, 436:221–231, 2021

    2021

  42. [50]

    Baddar, and Yong Man Ro

    Dae Hoe Kim, Wissam J. Baddar, and Yong Man Ro. Micro- expression recognition with expression-state constrained spatio- temporal feature representations. In Proceedings of the 24th ACM International Conference on Multimedia, MM ’16, page 382–386, New York, NY , USA, 2016. Asso...

    2016

  43. [51]

    Spatiotemporal recurrent convolutional networks for recognizing spontaneous micro-expressions

    Zhaoqiang Xia, Xiaopeng Hong, Xingyu Gao, Xiaoyi Feng, and Guoying Zhao. Spatiotemporal recurrent convolutional networks for recognizing spontaneous micro-expressions. IEEE Transactions on Multimedia, 22(3):626–640, 2020

    2020

  44. [52]

    Satish Kumar Reddy, Yashwanth Reddy Meedi- male, Murari Mandal, and Santosh Kumar Vipparthi

    Monu Verma, M. Satish Kumar Reddy, Yashwanth Reddy Meedi- male, Murari Mandal, and Santosh Kumar Vipparthi. Automer: Spa- tiotemporal neural architecture search for microexpression recogni- tion. IEEE Transactions on Neural Networks and Learning Systems, 33(11):6116–6128, 2022

    2022

  45. [53]

    Rmt: Retentive networks meet vision transformers

    Qihang Fan, Huaibo Huang, Mingrui Chen, Hongmin Liu, and Ran He. Rmt: Retentive networks meet vision transformers. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5641–5651, 2024

    2024

  46. [54]

    Transformers without normalization

    Jiachen Zhu, Xinlei Chen, Kaiming He, Yann LeCun, and Zhuang Liu. Transformers without normalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025

    2025

  47. [55]

    In defense of fully connected layers in visual representation trans- fer

    Chen-Lin Zhang, Jian-Hao Luo, Xiu-Shen Wei, and Jianxin Wu. In defense of fully connected layers in visual representation trans- fer. In Bing Zeng, Qingming Huang, Abdulmotaleb El Saddik, Hongliang Li, Shuqiang Jiang, and Xiaopeng Fan, editors, Advances in Multimedia Informati...

    2017

  48. [56]

    Micro-expression recognition based on facial graph representation learning and facial action unit fusion

    Ling Lei, Tong Chen, Shigang Li, and Jianfeng Li. Micro-expression recognition based on facial graph representation learning and facial action unit fusion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 1571–1580, June 2021

    2021

  49. [57]

    Mmnet: Muscle motion-guided network for micro-expression recognition

    Hanting Li, Mingzhe Sui, Zhaoqing Zhu, and Feng Zhao. Mmnet: Muscle motion-guided network for micro-expression recognition. In Proceedings of the 31th International Joint Conference on Artificial Intelligence, 2022

    2022

  50. [58]

    C3dbed: Facial micro- expression recognition with three-dimensional convolutional neural network embedding in transformer model

    Hang Pan, Lun Xie, and Zhiliang Wang. C3dbed: Facial micro- expression recognition with three-dimensional convolutional neural network embedding in transformer model. Engineering Applications of Artificial Intelligence, 123:106258, 2023

    2023

  51. [59]

    Fa- cialflownet: Advancing facial optical flow estimation with a diverse dataset and a decomposed model

    Jianzhi Lu, Ruian He, Shili Zhou, Weimin Tan, and Bo Yan. Fa- cialflownet: Advancing facial optical flow estimation with a diverse dataset and a decomposed model. In Proceedings of the 32nd ACM International Conference on Multimedia, MM ’24, page 2194–2203. ACM, October 2024

    2024

  52. [60]

    A multi-prior fusion network for video-based micro- expression recognition

    Chuang Ma, Shaokai Zhao, Yu Pei, Liang Xie, Erwei Yin, and Ye Yan. A multi-prior fusion network for video-based micro- expression recognition. In ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5, 2025

    2025

  53. [61]

    Facial 3d regional structural mo- tion representation using lightweight point cloud networks for micro- expression recognition

    Ren Zhang, Jianqin Yin, Chao Qi, Yonghao Dang, Zehao Wang, Zhicheng Zhang, and Huaping Liu. Facial 3d regional structural mo- tion representation using lightweight point cloud networks for micro- expression recognition. IEEE Transactions on Affective Computing, pages 1–15, 2025

    2025

  54. [62]

    Sfamnet: A scene flow attention-based micro-expression network

    Gen-Bing Liong, Sze-Teng Liong, Chee Seng Chan, and John See. Sfamnet: A scene flow attention-based micro-expression network. Neurocomputing, 566:126998, 2024

    2024

  55. [63]

    Adap- tive temporal motion guided graph convolution network for micro- expression recognition

    Fengyuan Zhang, Zhaopei Huang, Xinjie Zhang, and Qin Jin. Adap- tive temporal motion guided graph convolution network for micro- expression recognition. In 2024 IEEE International Conference on Multimedia and Expo (ICME), pages 1–6, 2024

    2024

  56. [64]

    A dual-branch approach with multi-stage semantic integration and dual optical flow for micro- expression recognition

    Shuhuan Zhao, Peijing Zhao, Zixin Hao, and Shuaiqi Liu. A dual-branch approach with multi-stage semantic integration and dual optical flow for micro- expression recognition. 2025. PREPRINT (Version 1) available at Research Square

    2025

  57. [65]

    A two-stage 3d cnn based learning method for spontaneous micro-expression recognition

    Sirui Zhao, Hanqing Tao, Yangsong Zhang, Tong Xu, Kun Zhang, Zhongkai Hao, et al. A two-stage 3d cnn based learning method for spontaneous micro-expression recognition. Neurocomputing, 448:276–289, 2021

    2021

  58. [66]

    A novel micro-expression recognition approach using attention-based magnification-adaptive networks

    Mengting Wei, Wenming Zheng, Yuan Zong, Xingxun Jiang, Cheng Lu, and Jiateng Liu. A novel micro-expression recognition approach using attention-based magnification-adaptive networks. In ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing...

    2022

  59. [67]

    Feature representation learning with adaptive displacement generation and transformer fusion for micro-expression recognition

    Zhijun Zhai, Jianhui Zhao, Chengjiang Long, Wenju Xu, Shuangjiang He, and Huijuan Zhao. Feature representation learning with adaptive displacement generation and transformer fusion for micro-expression recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision a...

    2023

  60. [68]

    A dual stream spatio-temporal deep network for micro-expression recognition us- ing upper facial features

    Nikin Matharaarachchi and Muhammad Fermi Pasha. A dual stream spatio-temporal deep network for micro-expression recognition us- ing upper facial features. Neural Computing and Applications, 37(3):1271–1287, November 2024

    2024

  61. [69]

    Boosting micro-expression recognition via self- expression reconstruction and memory contrastive learning

    Yongtang Bao, Chenxi Wu, Peng Zhang, Caifeng Shan, Yue Qi, and Xianye Ben. Boosting micro-expression recognition via self- expression reconstruction and memory contrastive learning. IEEE Transactions on Affective Computing, 15(4):2083–2096, 2024

    2024

  62. [70]

    Parallel spatiotemporal network to recognize micro-expression

    Jingting Li, Su-Jing Wang, Yong Wang, Haoliang Zhou, and Xiaolan Fu. Parallel spatiotemporal network to recognize micro-expression. Neurocomputing, 636:129891, 2025

    2025

  63. [71]

    Visualizing data using t-sne

    Laurens van der Maaten and Geoffrey Hinton. Visualizing data using t-sne. Journal of Machine Learning Research, 9(86):2579–2605, 2008

    2008

  64. [72]

    Selvaraju, Michael Cogswell, Abhishek Das, Ra- makrishna Vedantam, Devi Parikh, and Dhruv Batra

    Ramprasaath R. Selvaraju, Michael Cogswell, Abhishek Das, Ra- makrishna Vedantam, Devi Parikh, and Dhruv Batra. Grad-cam: Visual explanations from deep networks via gradient-based localiza- tion. In 2017 IEEE International Conference on Computer Vision (ICCV), pages 618–626, 2...

    2017

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.