Pith. sign in

REVIEW 3 major objections 5 minor 53 references

Learner Attentiveness and Engagement Analysis in Online Education Using Computer Vision

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A webcam-only vision system classifies four learner states and outputs a single live attentiveness index, with the EfficientNet variant reaching 80.32% engagement accuracy on DAiSEE.

desk verdict Solid applied DAiSEE classification work with a real pipeline, but the headline Attentiveness Index is fit on an undocumented private set with no validation, so the central claim overreaches. read the letter →

arxiv 2412.00429 v1 pith:HWLRJ3TJ submitted 2024-11-30 cs.CV cs.AI

Sharva Gogawale , Madhura Deshpande , Parteek Kumar , Irad Ben-Gal This is my paper · ORCID
classification cs.CVcs.AI
keywords attentivenessindexaffectivestatesengagementdetectiononlineeducationcomputervisionDAiSEEEfficientNetfocalloss
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

The paper claims that a webcam-only computer vision system can classify four learning-centered affective states—boredom, engagement, confusion, and frustration—from a learner's face in real time, then combine those classifications into one Attentiveness Index an instructor can watch during a live online class. It argues that the proposed hybrid CNN and EfficientNet models with focal loss outperform previous methods on the DAiSEE dataset, reporting 80.32% accuracy for engagement and similar or better figures for the other states. If the claim holds, instructors gain an inexpensive, deployable feedback channel that previously required expensive gaze trackers, sensor suites, or manual observation.

What carries the argument

The load-bearing object is the four parallel classifier branches, each a convolutional feature extractor (one lightweight 0.47M-parameter CNN, one EfficientNetB2 backbone with 31.7M parameters) with a 4-way softmax for one affective state, trained with categorical focal loss to counter DAiSEE's class imbalance. On top of the concatenated outputs sits a multiple-linear-regression Attentiveness Index, Eq. (4), which converts the four intensity vectors into a scalar. The index does the explanatory work: it is the piece that turns raw classification scores into a pedagogically usable number.

What would settle it

Collect a new, diverse set of online-classroom videos with independent instructor attentiveness ratings; compute the paper's Attentiveness Index from the trained classifier; and measure the rank correlation between predicted and human-rated attentiveness. If the correlation is near zero or the originally fitted weights do not transfer, the 'comprehensive index' claim fails even if the per-state classification accuracy is reproduced.

Watch

Extended reading notes

Core claim

The central claim is that attentiveness in online learning can be reduced to a weighted sum of four classifier outputs, with the weights learned from instructor ratings: $A_i = -0.598 \cdot B + 1.539 \cdot E + 0.334 \cdot C - 0.085 \cdot F$. The paper asserts this formula is comprehensive, aligns with cognitive-science literature on boredom, confusion, and frustration, and that the underlying four-branch classifier achieves higher accuracy than prior methods on DAiSEE, with the EfficientNet variant reaching 80.32% engagement accuracy. It further claims that the full pipeline—live webcam capture, Haar-cascade face cropping, grayscale 64x64 preprocessing, per-state softmax prediction, and index computation—runs in real time and provides graphical and tabular analytics plus alerts to instructors.

Load-bearing premise

The paper assumes that the instructor ratings used to set the weights in the Attentiveness Index are representative of all instructors and all online classrooms, even though the number of raters, the selection of videos, and rater agreement are not reported.

Editorial extensions

If this is right

  • Instructors can monitor a live class and receive alerts when the aggregate engagement drops below a threshold, pinpointing moments of disengagement.
  • Because the system uses only a standard webcam, it can be deployed on cloud servers without specialized hardware, making it accessible to ordinary online classrooms.
  • Per-state classifier outputs allow instructors to see spikes in confusion or frustration during specific lecture segments, not just a global attention score.
  • The system can analyze multiple lectures to recommend teaching patterns that sustain higher engagement, as the paper's pipeline description indicates.

Reading between the lines

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

  • Not argued in the paper: because the Attentiveness Index is fit to an unreported instructor-annotated subset, its claimed comprehensiveness should be tested by re-fitting on a large, diverse label set and checking out-of-sample agreement.
  • Not argued in the paper: frame-level accuracy may not equal video-level attentiveness; a natural extension is temporal smoothing over 10-second DAiSEE snippets and validation against continuous human ratings.
  • Not argued in the paper: deploying in real classrooms with varied lighting, glasses, masks, and camera angles would stress-test the Haar-cascade face crop and 64x64 grayscale preprocessing, which are fixed in this pipeline.
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

3 major / 5 minor

Summary. The paper proposes a computer-vision pipeline for real-time learner attentiveness analysis in online education. It trains two multi-branch architectures (a lightweight CNN and an EfficientNetB2 backbone) with focal loss on the DAiSEE dataset, predicting intensity levels for boredom, engagement, confusion, and frustration. On top of these predictions, the authors define an Attentiveness Index (Ai) in Eq. (4) as a weighted linear combination of the four affective states, with weights obtained by multiple linear regression on a privately annotated subset of the data. They report per-state classification accuracies (e.g., 80.32% for engagement) and claim improved attentiveness detection over prior work. They also describe an end-to-end web-based system for instructors. The classification accuracy claims are benchmarked against the public DAiSEE test set, but the Attentiveness Index validation and several statistical reporting details are incomplete.

Significance. If the classification results are reproducible and the Attentiveness Index is properly validated, the system could be a practical, low-cost tool for real-time engagement monitoring. The use of a public benchmark (DAiSEE) and comparison with several prior methods are strengths, as is the availability of a deployable pipeline. However, the central 'better attentiveness detection' claim depends on the Attentiveness Index, which is currently not validated with any independent data, and the accuracy comparisons lack per-class metrics and confidence intervals. The work is therefore of moderate significance until these gaps are addressed.

major comments (3)
  1. [Section 4, Eq. (4)] The Attentiveness Index is derived from a privately annotated 'subset of the dataset' using multiple linear regression, but the manuscript does not report the subset size, selection criteria, annotator agreement, regression fit quality, or any out-of-sample validation. It is also unspecified whether the regression used ground-truth labels or classifier predictions. Since the abstract's claim of 'better attentiveness detection' rests on this index as the summative metric, the current evidence does not substantiate that the index measures attentiveness. The authors should provide details of the annotation and regression validation, or temper the claim to reflect that the index is a proposed heuristic.
  2. [Section 6, Tables 2 and 3] The reported accuracies are point estimates on a highly imbalanced test set. As shown in Table 1, the engagement class has only 4 test videos for level 0, and similar extreme imbalances exist for other states. Overall accuracy is therefore dominated by majority classes and does not convey per-class performance. The authors should report per-class precision, recall, F1-score, and confidence intervals or other uncertainty measures. Without these, the claimed improvements over prior methods (e.g., 80.32% vs 63.59% in Table 3) are not statistically grounded.
  3. [Section 6 and Tables 2/3] The comparison with prior work does not confirm that all methods were evaluated under identical protocols. Prior methods may use different frame-level or video-level predictions, different pre-processing (e.g., face detection and frame discarding), or different class-weighting schemes. The manuscript should explicitly state the evaluation protocol used for the proposed models and verify that the cited accuracies for prior methods come from the same test split and metric definition, otherwise the 'better than state-of-the-art' conclusion is not reliable.
minor comments (5)
  1. [Section 3.5, Eq. (1)] The focal loss parameters alpha and gamma are defined in Eq. (1) but their values are never reported in the training description. Providing these values is essential for reproducibility.
  2. [Section 3.2] The preprocessing discards frames where Haar cascade face detection fails, but the number or proportion of discarded frames is not reported. This could bias the evaluation and should be quantified.
  3. [Table captions] Table captions contain a typo: 'T able 1' and 'T able 2' instead of 'Table 1' and 'Table 2'.
  4. [Section 5] The deployment link (https://smart-edu-system-01.herokuapp.com) may not be stable; consider providing a repository or local demo instructions.
  5. [Section 7] The authors' contributions list 'investigation, T.R.' but no author with those initials is listed; this appears to be an error.

Circularity Check

1 steps flagged · score 6.0 of 10

The Attentiveness Index in Eq. (4) is a regression fit to private instructor ratings, reported without out-of-sample validation, so the 'attentiveness' claims reduce to that fit while the per-state classification accuracies remain externally benchmarked.

  1. fitted input called prediction [Section 4, Eq. (4)]
    "A machine learning algorithm based on multiple linear regression was applied on top of this newly formed data subset provided the weights for each affective state, consequently providing the overall mathematical formula linking the affective states to a unique attentiveness index. The obtained mathematical formula is : Ai = −0.598 · B + 1.539 · E + 0.334 · C − 0.085 · F (4)"

    The index Ai is not independently derived or validated; it is the fitted output of a multiple linear regression on instructor-rated attentiveness scores for an unreported subset. By construction, the least-squares coefficients are optimized to reproduce those very ratings, so reporting Eq. (4) as a comprehensive measure and using it to substantiate better attentiveness detection presents the fit itself as the evidence. The paper gives no held-out evaluation, no subset statistics, no inter-annotator agreement, and no comparison against an external attentiveness criterion; the only validation is qualitative consistency with cognitive-science references, which does not test the fitted weights.

full rationale

The classification results are not circular: the hybrid CNN and EfficientNet models are trained and evaluated on the public DAiSEE train/test split, and the reported accuracies (e.g., 80.32% engagement for EfficientNet with focal loss) are externally grounded against prior published results on the same benchmark. The self-citations to prior work by co-author Parteek Kumar (refs. [8], [53]) appear only as related work and do not carry any load-bearing argument, so they do not raise the circularity score. The circularity concern is confined to the Attentiveness Index. Section 4 derives Eq. (4) by multiple linear regression on a private, undocumented subset of DAiSEE videos that instructors rated for attentiveness. The resulting coefficients are then presented as a validated mathematical formula and the abstract claims better attentiveness detection than state-of-the-art methods, even though the comparison tables only cover per-state affective classification. The index is a fit to the very target construct it is claimed to measure, with no independent derivation, no reported regression diagnostics, and no out-of-sample test. Thus the paper's central attentiveness-specific claim reduces to a fitted regression on an unpublished annotation set, while the per-state classification accuracy remains an independent, non-circular contribution. This warrants a partial circularity score of 6 rather than a lower score: the main classifier benchmark is sound, but the headline attentiveness metric is a fitted input renamed as a discovered index.

Assumptions & free parameters 3 free parameters · 5 assumptions · 1 invented entities

The central claim rests on standard supervised learning plus a private annotation fit. The regression weights and focal loss hyperparameters are free parameters; the Attentiveness Index is a constructed metric without external validation.

free parameters (3)
  • Attentiveness index regression weights = -0.598 (B), 1.539 (E), 0.334 (C), -0.085 (F)
    Obtained from multiple linear regression on a private instructor-rated subset; no confidence intervals, R-squared, or validation reported.
  • Focal loss alpha and gamma = not reported
    Eq. (1) depends on alpha and gamma, but the paper never states the values used for training.
  • Initial learning rate = 0.001
    Adam optimizer learning rate is a hand-chosen hyperparameter that affects the reported accuracy, with no tuning analysis.
assumptions (5)
  • domain assumption DAiSEE crowd-sourced labels with expert-correlated gold standard are valid ground truth for boredom, engagement, confusion, and frustration.
    The entire supervised training and evaluation treats these labels as correct; the paper does not audit label noise or annotator bias.
  • domain assumption Facial appearance alone, after Haar-cascade cropping and 64x64 grayscale conversion, carries enough signal to infer attentiveness.
    The pipeline discards audio, body pose, gaze depth, and context; this assumes visual face cues are sufficient for the attentiveness index.
  • ad hoc to paper Instructor ratings on a 1 to 10 scale are reliable, internally consistent, and linearly related to the four affective-state intensity labels.
    Section 4 fits a linear regression but reports no inter-rater reliability, rating distribution, or model diagnostics; linearity is assumed.
  • standard math Multiple linear regression via least squares is an appropriate model for mapping affective states to attentiveness.
    This is a standard statistical tool; the paper provides no evidence that the linear model fits the annotation data well.
  • ad hoc to paper The four affective-state intensities combine additively, with no interaction terms, to yield attentiveness.
    Eq. (4) contains only main effects; no interactions or nonlinear terms are considered or tested.
invented entities (1)
  • Attentiveness Index (Ai)
    purpose: Single scalar summary of learner attentiveness from four affective-state intensities for instructor dashboards.
    It is defined by Eq. (4) with weights fit to a private annotation subset; no external behavioral or learning-outcome validation is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learner Attentiveness and Engagement Analysis in Online Education Using Computer Vision." pith.science (2026). https://pith.science/paper/HWLRJ3TJ

@misc{pith2026241200429,
  author       = {Pith},
  title        = {Pith review of: Learner Attentiveness and Engagement Analysis in Online Education Using Computer Vision},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HWLRJ3TJ}},
  note         = {Machine review of arXiv:2412.00429}
}
read the original abstract

In recent times, online education and the usage of video-conferencing platforms have experienced massive growth. Due to the limited scope of a virtual classroom, it may become difficult for instructors to analyze learners' attention and comprehension in real time while teaching. In the digital mode of education, it would be beneficial for instructors to have an automated feedback mechanism to be informed regarding learners' attentiveness at any given time. This research presents a novel computer vision-based approach to analyze and quantify learners' attentiveness, engagement, and other affective states within online learning scenarios. This work presents the development of a multiclass multioutput classification method using convolutional neural networks on a publicly available dataset - DAiSEE. A machine learning-based algorithm is developed on top of the classification model that outputs a comprehensive attentiveness index of the learners. Furthermore, an end-to-end pipeline is proposed through which learners' live video feed is processed, providing detailed attentiveness analytics of the learners to the instructors. By comparing the experimental outcomes of the proposed method against those of previous methods, it is demonstrated that the proposed method exhibits better attentiveness detection than state-of-the-art methods. The proposed system is a comprehensive, practical, and real-time solution that is deployable and easy to use. The experimental results also demonstrate the system's efficiency in gauging learners' attentiveness.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 50 canonical work pages

  1. [1]

    and Bej, T

    Kundu, A. and Bej, T. (2021). ”COVID-19 response: students’ readiness for shifting classes online”, Corporate Governance, Vol. 21 No. 6, pp. 1250-1270

    work page 2021

  2. [2]

    and Graesser, A.C

    D’Mello, S.K., Craig, S.D. and Graesser, A.C. (2009). MultiMethod Assessment of Affective Experience and Expression during Deep Learning.International Journal of Learning Technology, 4(3/4), 165–187

    work page 2009

  3. [3]

    & Doleck, T

    Basnet, R.B., Johnson, C. & Doleck, T. (2022). Dropout prediction in Moocs using deep learning and machine learning. Educ Inf Technol , 27, 11499–11513. https://doi.org/10.1007/s10639-022-11068-7 17

    work page doi:10.1007/s10639-022-11068-7 2022

  4. [4]

    Liu, Y., Chen, J., Zhang, M. et al. Student engagement study based on multi-cue detection and recognition in an intelligent learning environment. Multimed Tools Appl 77, 28749–28775 (2018). https://doi.org/10.1007/s11042-018-6017-2

    work page doi:10.1007/s11042-018-6017-2 2018

  5. [5]

    Ouyang, F., Wu, M., Zheng, L. et al. (2023). Integration of artificial intel- ligence performance prediction and learning analytics to improve student learning in online engineering course. Int J Educ Technol High Educ , 20, 4. https://doi.org/10.1186/s41239-022-00372-4

    work page doi:10.1186/s41239-022-00372-4 2023

  6. [6]

    Karamimehr, M

    Z. Karamimehr, M. M. Sepehri, and S. Sibdari (2020). Automatic method to identify e-learner emotions using behavioral cues. IEEE Transactions on Learning Technologies, 13(4), 762–776

    work page 2020

  7. [7]

    Zheng, S

    X. Zheng, S. Hasegawa, M.-T. Tran, K. Ota, and T. Unoki (2021). Estimation of learners’ engagement using face and body features by transfer learning. In Artificial Intelligence in HCI , pp. 541–552. Springer International Publishing

    work page 2021

  8. [8]

    & Tekchandani, R

    Gupta, S., Kumar, P. & Tekchandani, R. Artificial intelligence based cognitive state prediction in an e-learning environment using multimodal data. Multimed Tools Appl (2024). https://doi.org/10.1007/s11042-023-18021-x

    work page doi:10.1007/s11042-023-18021-x 2024

Show all 53 references
  1. [9]

    Yang, C.-Y

    F.-Y. Yang, C.-Y. Chang, W.-R. Chien, Y.-T. Chien, and Y.-H. Tseng (2013). Tracking learners’ visual attention during a multimedia presentation in a real classroom. Computers & Education , 62, 208–220

    2013

  2. [10]

    G. Y.-M. Kao, X.-Z. Chiang, and T. Foulsham (2019). Reading behavior and the effect of embedded selfies in role-playing picture e-books: An eye-tracking investigation. Computers & Education , 136, 99–112

    2019

  3. [11]

    Ishikawa, T., Baker, S., Matthews, I., & Kanade, T. (2004). Passive driver gaze tracking with active appearance models. In Proc. World Congress on Intelligent Transportation Systems

    2004

  4. [12]

    Wood, E., Baltrusaitis, T., Morency, L.-P., Robinson, P., & Bulling, A. (2016). A 3D morphable eye region model for gaze estimation. In European Conference on Computer Vision , 297–313

    2016

  5. [13]

    Wood, E., & Bulling, A. (2014). Eyetab: Model-based gaze estimation on unmodi- fied tablet computers. In Symposium on Eye Tracking Research and Applications, 207–210

    2014

  6. [14]

    Valenti, R., Sebe, N., & Gevers, T. (2011). Combining head pose and eye location information for gaze estimation. IEEE Transactions on Image Processing , 21(2), 802–815

    2011

  7. [15]

    Dewan, M. A. A., Murshed, M., & Lin, F. (2019). Engagement detection in online learning: A review. Smart Learning Environments, 6(1), 1–20. 18

    2019

  8. [16]

    Dhall, A., Sharma, G., Goecke, R., & Gedeon, T. (2020). Emotiw 2020: Driver gaze, group emotion, student engagement and physiological signal based chal- lenges. In Proceedings of the 2020 International Conference on Multimodal Interaction, 784–789

    2020

  9. [17]

    Wu, J., Yang, B., Wang, Y., & Hattori, G. (2020). Advanced multi-instance learning method with multi-features engineering and conservative optimization for engagement intensity prediction. In Proceedings of the 2020 International Conference on Multimodal Interaction , 777–783

    2020

  10. [18]

    E., & Boncelet, C

    Zhu, B., Lan, X., Guo, X., Barner, K. E., & Boncelet, C. (2020). Multi-rate attention based GRU model for engagement prediction. InProceedings of the 2020 International Conference on Multimodal Interaction , 841–848

    2020

  11. [19]

    F., Wiggins, J

    Grafsgaard, J. F., Wiggins, J. B., Boyer, K. E., Wiebe, E. N., & Lester, J. C. (2013). Automatically recognizing facial expression: Predicting engagement and frustration. In International Conference on Educational Data Mining , 2013

    2013

  12. [20]

    Whitehill, J., Serpell, Z., Lin, Y.-C., Foster, A., & Movellan, J. R. (2014). The faces of engagement: Automatic recognition of student engagement from facial expressions. IEEE Transactions on Affective Computing , 5(1), 86–98

    2014

  13. [21]

    Zaletelj, J., & Kosir, A. (2017). Predicting students’ attention in the classroom from Kinect facial and body features. EURASIP Journal on Image and Video Processing, 2017(80)

    2017

  14. [22]

    Whitehill, J. (2014). The faces of engagement: Automatic recognition of student engagement from facial expressions. IEEE Transactions on Affective Computing , 5(1), 86–98

    2014

  15. [23]

    Yang, J., Wang, K., & Qiao, Y. (2018). Deep recurrent multi-instance learning with spatio-temporal features for engagement intensity prediction. In Proceedings of the 20th ACM International Conference on Multimodal Interaction . ACM

    2018

  16. [24]

    Niu, X., Han, H., & Chen, X. (2018). Automatic engagement prediction with gap feature. In Proceedings of the 20th ACM International Conference on Multimodal Interaction. ACM

    2018

  17. [25]

    Monkaresi, H., & Bosch, N. B. (2017). Automated detection of engagement using video-based estimation of facial expressions and heart rate. IEEE Transactions on Affective Computing , 8(1), 15–28

    2017

  18. [26]

    Carreira, J., & Zisserman, A. (2017). Quo vadis, action recognition? a new model and the kinetics dataset. In Proc. CVPR

    2017

  19. [27]

    Krithika, L. B. (2016). Student emotion recognition system (SERS) for e-learning improvement based on learner concentration metric. Procedia Computer Science, 19 85, 767–776

    2016

  20. [28]

    R., Khanal, T

    Sharma, P., Eseng¨ on¨ ul, M., Khanal, S. R., Khanal, T. T., Filipe, V., & Reis, M. J. C. S. (2018). Student Concentration Evaluation Index in an E-learning Con- text Using Facial Emotion Analysis. In Technology and Innovation in Learning, Teaching and Education, TECH-EDU 2018...

    2018

  21. [29]

    Gupta, A., D’Cunha, A., Awasthi, K., & Balasubramanian, V. (2016). Daisee: Towards user engagement recognition in the wild. arXiv preprint arXiv:1609.01885

    2016 arXiv

  22. [30]

    Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition , 1–9

    2015

  23. [31]

    Tran, D., Bourdev, L., Fergus, R., Torresani, L., & Paluri, M. (2015). Learning spatiotemporal features with 3D convolutional networks. In Proceedings of the IEEE International Conference on Computer Vision , 4489–4497

    2015

  24. [32]

    A., Guadarrama, S., Rohrbach, M., Venugopalan, S., Saenko, K., & Darrell, T

    Donahue, J., Hendricks, L. A., Guadarrama, S., Rohrbach, M., Venugopalan, S., Saenko, K., & Darrell, T. (2015). Long-term recurrent convolutional networks for visual recognition and description. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition ,...

    2015

  25. [33]

    Geng, L., Xu, M., Wei, Z., & Zhou, X. (2019). Learning deep spatiotemporal features for engagement recognition of online courses. In 2019 IEEE Symposium Series on Computational Intelligence (SSCI) , 442–447

    2019

  26. [34]

    Zhang, H., Xiao, X., Huang, T., Liu, S., Xia, Y., & Li, J. (2019). A novel end-to-end network for automatic student engagement recognition. In 2019 IEEE 9th International Conference on Electronics Information and Emergency Communication (ICEIEC), 342–345

    2019

  27. [35]

    Liao, J., Liang, Y., & Pan, J. (2021). Deep facial spatiotemporal network for engagement prediction in online learning. Applied Intelligence, 1–13

    2021

  28. [36]

    Dewan, M. A. A., Lin, F., Wen, D., Murshed, M., & Ud-din, Z. (2018). A deep learning approach to detecting engagement of online learners. In 2018 IEEE SmartWorld, Ubiquitous Intelligence & Computing, Advanced & Trusted Computing, Scalable Computing & Communications, Cloud & Bi...

    2018

  29. [37]

    Murshed, M., Dewan, M. A. A., Lin, F., & Wen, D. (2019). Engagement detection in e-learning environments using convolutional neural networks. In 2019 IEEE 20 Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Per- vasive Intelligence and Computing, Intl Conf...

    2019

  30. [38]

    A., & Person, N

    Lehman, B. A., & Person, N. (2009). Understanding students’ affective states dur- ing learning. In Ninth International Conference on Intelligent Tutoring Systems, ITS 2008, LNCS 5091

    2009

  31. [39]

    Gupta, A., D’Cunha, A., Awasthi, K., & Balasubramanian, V. (2016). Daisee: Towards user engagement recognition in the wild

    2016

  32. [40]

    A., Gal, K., Segal, A., Sripathi, K., Kim, H

    Geller, S. A., Gal, K., Segal, A., Sripathi, K., Kim, H. G., Facciotti, M. T., Igo, M., Hoernle, N., & Karger, D. (2021). New methods for confusion detection in course forums: Student, teacher, and machine. IEEE Transactions on Learning Technologies, 14(5), 665–679

    2021

  33. [41]

    d., D’Mello, S.K., Rodrigo, M.M.T

    Baker, R.S.J. d., D’Mello, S.K., Rodrigo, M.M.T. and Graesser, A.C. (2010). Better to Be Frustrated than Bored: The Incidence, Persistence, and Impact of Learners’ Cognitive-Affective States during Interactions with Three Differ- ent Computer-Based Learning Environments. Inter...

    2010

  34. [42]

    D’Mello, S.K., Lehman, B., Pekrun, R., & Graesser, A.C. (2014). Confusion Can Be Beneficial for Learning. Learning & Instruction

    2014

  35. [43]

    Leite, I., McCoy, M., Ullman, D., Salomons, N., & Scassellati, B. (2015). Compar- ing models of disengagement in individual and group interactions. International Conference on Human-Robot Interaction

    2015

  36. [44]

    Kapoor, A., & Picard, R. W. (2005). Multimodal affect recognition in learning environments. ACM International Conference on Multimedia , 677–682

    2005

  37. [45]

    H., & Venables, L

    Fairclough, S. H., & Venables, L. (2006). Prediction of subjective states from psy- chophysiology: A multivariate approach. Biological Psychology, 71(1), 100–110

    2006

  38. [46]

    Chaouachi, M., & Frasson, C. (2010). Exploring the relationship between learner EEG mental engagement and affect. In Intelligent Tutoring Systems , 291–293

    2010

  39. [47]

    A., & D’Mello, S

    Monkaresi, H., Bosch, N., Calvo, R. A., & D’Mello, S. K. (2017). Automated Detection of Engagement Using Video-Based Estimation of Facial Expressions and Heart Rate. IEEE Transactions on Affective Computing , 8(1), 15–28

    2017

  40. [48]

    Tan, M., & Le, Q. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In Proc. Int. Conf. Mach. Learn. , 688, 6105–6114. 21

    2019

  41. [49]

    Lin, T.-Y., Goyal, P., Girshick, R., He, K., & Doll´ ar, P. (2020). Focal Loss for Dense Object Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42(2), 318–327

    2020

  42. [50]

    Huang, T., Mei, Y., Zhang, H., Liu, S., & Yang, H. (2019). Fine-grained Engage- ment Recognition in Online Learning Environment. In 2019 IEEE 9th Interna- tional Conference on Electronics Information and Emergency Communication (ICEIEC), 338–341

    2019

  43. [51]

    K., Prasad, S

    Mehta, N. K., Prasad, S. S., Saurav, S., Saini, R., & Singh, S. (2022). Three dimensional DenseNet self-attention neural network for automatic detection of student’s engagement. Appl. Intell. , 52, 1–21

    2022

  44. [52]

    Mandia, S., Singh, K., & Mitharwal, R. (2022). Vision Transformer for Automatic Student Engagement Estimation. In 2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS) , Genova, Italy, 1–6. doi: 10.1109/IPAS55744.2022.10052945

    2022

  45. [53]

    A vision-based multi-cues approach for individual stu- dents’ and overall class engagement monitoring in smart classroom environments

    Pabba, C., Kumar, P. A vision-based multi-cues approach for individual stu- dents’ and overall class engagement monitoring in smart classroom environments. Multimed Tools Appl 83, 52621–52652 (2024). https://doi.org/10.1007/s11042- 023-17533-w 22

    2024 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.