Pith. sign in

REVIEW 4 major objections 8 minor 38 references

Development and evaluation of a deep learning algorithm for German word recognition from lip movements

T0 review · 4 major / 8 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read This paper reports a deep-learning system that recognizes 18 German words from lip movements, reaching 87.3% accuracy on known speakers and 62.61% on unseen speakers.

desk verdict First German lipreading dataset and benchmark, but the headline accuracies rest on unverified ASR timestamps and no released artifacts, so treat the numbers as provisional. read the letter →

arxiv 2504.15792 v1 pith:BSA5VMBH submitted 2025-04-22 cs.CV

Dinh Nam Pham , Torsten Rahne This is my paper · ORCID
classification cs.CV
keywords lipreadingvisualspeechrecognitionGermanlanguagedeeplearninggatedrecurrentunit3DconvolutionalneuralnetworkwordclassificationGRUConv
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 sets out to build the first automatic deep-learning lip-reading system for German. It claims that a hybrid network called GRUConv, combining 3D convolutions with gated recurrent units, can classify 18 visually distinguishable German words from short lip-movement videos, reaching 87.3% accuracy on speakers it has seen and 62.61% on speakers it has not. The result matters because no comparable German-language system existed, and because it suggests German lip-reading can reach the accuracy level reported for English systems. The authors also show that cropping to the mouth region matters far more than color space: lip crops give about 70% validation accuracy versus 34% for face crops, while color spaces differ by only a few percent.

What carries the argument

The load-bearing component is the GRUConv model, a hybrid in which three 3D-convolutional blocks (with batch normalization, max pooling, and dropout) extract spatial appearance at each frame, two bidirectional GRU layers track how the mouth region changes over the 28-frame input, and a multilayer perceptron maps the resulting representation to 18 word classes via softmax. The mouth-region crop, resized to 150×100 pixels and converted to LAB color space, supplies the input; the fixed 28-frame window is padded by repeating the last frame. This architecture is what the comparison experiments identify as the highest-accuracy configuration, reaching 77.59% validation accuracy on dataset A and then 87.3% on dataset B's test set.

What would settle it

Hand-correct the word labels and timestamps of a random sample of the training segments, retrain GRUConv from scratch on the corrected subset, and evaluate on datasets B and C; a large change in accuracy or confusion pattern would show that the reported numbers depend on the automatic labeling. In parallel, test the model on a manually verified set of German words longer than 28 frames; if accuracy drops sharply on those, the fixed frame window is truncating information.

Watch

Extended reading notes

Core claim

On the authors' own terms, the central discovery is that a GRUConv architecture—3D convolutional layers for spatial features followed by two bidirectional gated recurrent unit layers for temporal features, then a softmax classifier—can recognize German words from lip movements at accuracy comparable to English-language lip-reading systems. Trained on 38,391 automatically segmented clips of 18 words from 32 speakers, the model reaches 87.3% correct classification on a test set from known speakers and 62.61% on a test set of unseen speakers. The paper further reports that the largest accuracy gains come from the choice of input region rather than color space or model family, and that expanding the training data improved accuracy more than changing architecture.

Load-bearing premise

The training labels and word timestamps come from an automatic speech recognizer and were not manually verified; if they mislabel or mistime words, the network learns from noisy targets and the reported accuracies no longer describe clean German lip-reading.

Editorial extensions

If this is right

  • A German-language machine lip-reading system is feasible at word level, with accuracy on known speakers around 87% and on unseen speakers around 63%.
  • The mouth-region crop, not the color space, is the decisive preprocessing choice; face crops lose roughly half the accuracy.
  • Adding training data appears to improve accuracy more than changing model architecture, so dataset construction is the main scaling lever.
  • The same pipeline could extend to more German word classes and, with future changes such as connectionist temporal classification, to variable-length words or sentences.
  • If the speaker-independent result holds, the system could support automated subtitles for hearing-impaired viewers and silent dictation in noisy environments.

Reading between the lines

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

  • Because only validation and test clips were manually cleaned, a fully hand-cleaned training set might shift the reported accuracies; the direction and size of that shift is unknown from the paper alone.
  • The fixed 28-frame input suggests that longer words are compressed or truncated; variable-length inputs with connectionist temporal classification, which the authors mention as future work, would likely improve unseen-speaker accuracy if word-length variation is the bottleneck.
  • The claim that dataset size matters more than model choice is a testable statement: training GRUConv on progressively larger random subsets of Dataset B should show a smooth accuracy curve that flattens, revealing the marginal value of more German lip-reading data.
  • Because the videos come from natural, in-the-wild German speech, the method may transfer to real-world captioning or silent-dictation applications, but measuring that transfer requires testing on non-YouTube recordings with different lighting and camera angles.
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 / 8 minor

Summary. The manuscript presents a deep-learning system for recognizing 18 German words from silent lip movements, based on a newly assembled dataset of 38,391 video clips from 32 YouTube speakers. The authors compare three architectures (Conv3D, GRU, GRUConv), two crops (face vs. lips), and six color spaces, selecting hyperparameters by validation accuracy. They report 70.29% validation accuracy for the Conv3D model with lip crops on Dataset A, 77.59% for GRUConv on the same validation split, and, after retraining on the larger Dataset B, 87.3% test accuracy on known speakers and 62.61% on an unseen-speaker Dataset C. The paper claims this is the first automatic lip-reading system for German and that its accuracy is comparable to English-language systems.

Significance. If the reported numbers are robust, this is a useful contribution: it introduces the first German-language lip-reading dataset and demonstrates that a CNN-GRU hybrid can generalize to unseen speakers at 62.61% accuracy on 18 words. The experimental design has real strengths: speaker-disjoint datasets, a held-out test set, validation-based model selection, and explicit comparisons of preprocessing choices. The paper is also transparent about its main limitations, including automatic speech-recognition-based labeling and fixed-length video processing. However, the headline accuracies depend on the automatic word boundaries produced by Vosk, and the manuscript does not provide evidence that these boundaries are reliable enough to rule out segmentation artifacts. The lack of repeated runs and confidence intervals further limits the strength of the claims, especially for the small unseen-speaker set.

major comments (4)
  1. [2.1] The training data are labeled entirely by the Vosk speech recognizer; manual review is applied only to the validation and test sets and only to remove clips with incorrect words, not to check word boundaries. If Vosk has systematic timing biases (e.g., a consistent lead or lag for particular words), those biased clips are used in both training and testing, and the network could learn to exploit the segmentation artifact rather than the lip movement. This is load-bearing for the headline numbers (87.3% on Dataset B, 62.61% on Dataset C). Please quantify the reliability of the automatic boundaries, for example by manually verifying a stratified sample of training clips or by re-evaluating the final model on a test set with manually corrected boundaries, and report the resulting accuracy change.
  2. [2.2] All clips are forced to 28 frames by 'shortened or extended ... by repeating the last frame'. This means words longer than about one second are truncated, and the truncation point is determined by Vosk's timestamps. The combination of automatic boundaries and fixed-length truncation can create a non-linguistic cue (e.g., the model may see the same consistent portion of an adjacent phoneme). Please report the per-class duration distribution, the fraction of clips truncated or padded, and either adopt a variable-length model (CTC is mentioned as future work) or provide an analysis showing that truncation behavior does not differ systematically across classes.
  3. [3] All accuracy values are single runs with best-epoch selection on a validation set; no confidence intervals, repeated runs, or significance tests are reported. The conclusion that color spaces 'did not reveal any relevant different correct classification rates' is based on one-run differences of up to about 4 percentage points (LAB 73.47% vs. HSV 69.23%), which could easily be noise. Please report repeated-seed means and standard deviations for the key comparisons (crop, color space, model, final test) and use a paired test (e.g., McNemar) for the final evaluations.
  4. [3] Speaker-independent generalization is evaluated on Dataset C with only 4 speakers and 3,950 clips, yet only the aggregate accuracy of 62.61% is reported. With such a small number of speakers, the aggregate could be dominated by one or two speakers. Please report per-speaker accuracies and clip counts, and provide a confidence interval for the 62.61% estimate.
minor comments (8)
  1. [Table 1] The test split for Dataset B is not listed in the table, although the text and the 87.3% result imply that such a split exists; please state the number of test clips used.
  2. [Table 1] The word 'Valdiation' in Table 1 should be 'Validation'.
  3. [Table 2] The word 'bedeudet' in Table 2 appears to be a typo for 'bedeutet' (the German word for 'means').
  4. [Abstract and 2.1] The abstract says 1,806 video clips were 'split into word segments'; please clarify that these are 1,806 source videos from which 38,391 word-segment clips were generated by automatic speech recognition.
  5. [Figure 5] The confusion matrices described in the text are not visible in this version of the manuscript; please include the figure or provide a table of per-class accuracies so the claimed confusions can be examined.
  6. [4] The Discussion contains a typo: 'decrreased' should be 'decreased'.
  7. [4] The comparison with English-language systems (word accuracies of 76%-98%) is not controlled, because the word sets, number of classes, and datasets differ; please qualify the statement that the German system is 'comparable to English-language algorithms'.
  8. [4] The claim that 'dataset size impacts accuracy more than model choice' is not directly supported, since Datasets A and B differ not only in size but also in speakers and in the exact training procedure; please soften the claim or add a controlled comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: an empirical measurement paper with held-out test evaluation and no self-citation chain or definitional reduction.

full rationale

This paper is an empirical measurement study, not a derivation. The central claims are the measured accuracies (87.3% on Dataset B test, 62.61% on Dataset C unseen speakers), which are computed by classifying held-out video clips and comparing the predicted word class with the assigned label. There is no equation in which a predicted quantity is defined in terms of a fitted parameter, no parameter is fitted to the test set, and no result is imported from the authors' prior work. The model-selection procedure (choosing crop, color space, and architecture on Dataset A validation, then training the final model on Dataset B and evaluating on Dataset B's test set and Dataset C) is standard and does not make the test accuracy circular. The only plausible concern is that the Vosk speech recognizer generated the word labels and timestamps for all clips, with manual review applied only to validation and test clips. That is a data-quality and external-validity limitation, and the paper itself acknowledges it in the Discussion ('Potential errors in dataset creation (speech recognition or automatic cropping) might have influenced results'). It is not a circularity reduction: the model's output is not defined as the Vosk label, the training objective is to predict a target distribution that the model can fail to learn, and the validation/test labels were manually verified for word identity. The shared segmentation pipeline between training and test could produce systematic timestamp artifacts, but that is a benchmark-construction risk, not a case where a 'prediction' is equivalent to its input by construction. No self-citation is load-bearing, no uniqueness theorem is imported, and no known result is renamed. The reported accuracy is an honest held-out measurement under the paper's stated labeling procedure.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The central accuracy claims rest primarily on the automatically labeled dataset and the fixed preprocessing choices. The most consequential free parameters are the image crop, color space, model architecture, and epoch, which were all selected based on validation performance on Dataset A. The listed axioms capture the unverified assumptions about label correctness, temporal alignment, and task design that the reported accuracies depend on.

free parameters (8)
  • image crop region = lips (150x100)
    Chosen after comparing face versus lip cropping on Dataset A validation; lip cropping improved accuracy from 34% to 70%.
  • color space = LAB
    Chosen based on highest validation accuracy (73.47%) among RGB, grayscale, HSV, LAB, XYZ, YCbCr on Dataset A; differences were small.
  • model architecture = GRUConv
    Chosen based on highest validation accuracy (77.59%) on Dataset A compared with Conv3D (73.47%) and GRU (60.34%).
  • best epoch = varying (e.g., 4297 for GRUConv on A, 1868 on B)
    Model parameters from the epoch with the highest validation accuracy were retained, introducing validation-set-dependent selection.
  • learning rate = 1e-6
    Fixed ad hoc for all models.
  • batch size = 128 training, 32 validation
    Fixed ad hoc; no sensitivity analysis.
  • number of frames = 28
    Videos were uniformly shortened or extended to 28 frames by repeating the last frame; choice not justified.
  • Conv3D/GRU hyperparameters = kernels 8/16/32, kernel size (3,5,5), pooling (1,2,2), dropout 0.5
    Chosen without reported sensitivity analysis or search; only one configuration per model family was tested.
assumptions (5)
  • domain assumption Vosk automatic speech recognition produces correct word labels and timestamps for training videos
    Section 2.1: word classes were assigned using Vosk API; only validation and test sets were manually reviewed, so training labels inherit ASR errors.
  • domain assumption The 28-frame window captures the full articulation of each word
    Section 2.2: all videos were uniformly shortened or extended to 28 frames by repeating the last frame; if a word's articulation exceeds 28 frames or boundaries are inaccurate, the model sees partial mouth movements.
  • standard math Neural networks with sufficient capacity can learn lip-reading from the provided data
    Section 1 cites universal approximation results; this is a background assumption for the whole deep-learning approach.
  • ad hoc to paper The 18 selected words are visually distinguishable via lip movements
    Section 2.1 selects multisyllabic, visually distinguishable words; this is a task-design assumption and limits generalization to the broader German lexicon.
  • domain assumption Automatic face and mouth detection (face_recognition, dlib) correctly localizes the mouth across all videos
    Section 2.2 uses these tools to crop the lip region; errors in detection would feed incorrect inputs to the model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Development and evaluation of a deep learning algorithm for German word recognition from lip movements." pith.science (2026). https://pith.science/paper/BSA5VMBH

@misc{pith2026250415792,
  author       = {Pith},
  title        = {Pith review of: Development and evaluation of a deep learning algorithm for German word recognition from lip movements},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BSA5VMBH}},
  note         = {Machine review of arXiv:2504.15792}
}
read the original abstract

When reading lips, many people benefit from additional visual information from the lip movements of the speaker, which is, however, very error prone. Algorithms for lip reading with artificial intelligence based on artificial neural networks significantly improve word recognition but are not available for the German language. A total of 1806 video clips with only one German-speaking person each were selected, split into word segments, and assigned to word classes using speech-recognition software. In 38,391 video segments with 32 speakers, 18 polysyllabic, visually distinguishable words were used to train and validate a neural network. The 3D Convolutional Neural Network and Gated Recurrent Units models and a combination of both models (GRUConv) were compared, as were different image sections and color spaces of the videos. The accuracy was determined in 5000 training epochs. Comparison of the color spaces did not reveal any relevant different correct classification rates in the range from 69% to 72%. With a cut to the lips, a significantly higher accuracy of 70% was achieved than when cut to the entire speaker's face (34%). With the GRUConv model, the maximum accuracies were 87% with known speakers and 63% in the validation with unknown speakers. The neural network for lip reading, which was first developed for the German language, shows a very high level of accuracy, comparable to English-language algorithms. It works with unknown speakers as well and can be generalized with more word classes.

Figures

Figures reproduced from arXiv: 2504.15792 by the authors.

Figure 1
Figure 1. Example of a video which was cropped to the face (red) and lips (blue). [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Schema of the GRUConv model. During training of each of the models, multiple epochs were iterated. In each epoch, the model was trained with each video in the training set with a learning rate of 10−6 , so that the learnable parameters of the neural network were updated. ADAM [28] was used as the optimization algorithm while cross entropy served as the loss function. Subsequently, each video in the validation set wa… view at source ↗
Figure 3
Figure 3. Schema of the training strategy. 2.4 Implementation The creation of the datasets, the processing pipeline, the models and the training process were implemented in the programming language Python. First, the YouTube links were manually collected for the dataset. Using youtube-dl [29], the videos were subsequently downloaded. For speech recognition of the videos, the Python module of the Vosk API [24] was used, which … view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Word recognition accuracy (correct classification rate) versus number of training epochs comparing (a) image [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Confusion matrices of predicted versus true word classes for (a) Dataset A and (b) Dataset B using GRUConv. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 32 canonical work pages

  1. [1]

    von Gablenz, E

    P. von Gablenz, E. Hoffmann, and I. Holube. Prävalenz von schwerhörigkeit in nord- und süddeutschland. HNO, 65(8):663–670, March 2017

    work page 2017

  2. [2]

    A survey on mouth modeling and analysis for sign language recognition

    Epameinondas Antonakos, Anastasios Roussos, and Stefanos Zafeiriou. A survey on mouth modeling and analysis for sign language recognition. 2015 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG), 1:1–7, 2015

    work page 2015

  3. [3]

    Comparison of image transform-based features for visual speech recognition in clean and corrupted videos

    Rowan Seymour, Darryl Stewart, and Ji Ming. Comparison of image transform-based features for visual speech recognition in clean and corrupted videos. EURASIP Journal on Image and Video Processing , 2008:1–9, 2008

    work page 2008

  4. [4]

    Audiovisual automatic speech recognition: Progress and challenges

    Gerasimos Potamianos. Audiovisual automatic speech recognition: Progress and challenges. The Journal of the Acoustical Society of America , 123(5 Supplement):3939–3939, May 2008

    work page 2008

  5. [5]

    A review of recent advances in visual speech decoding

    Ziheng Zhou, Guoying Zhao, Xiaopeng Hong, and Matti Pietikäinen. A review of recent advances in visual speech decoding. Image and Vision Computing, 32(9):590–605, September 2014

    work page 2014

  6. [6]

    https://landesdolmetscherzentrale-gebaerdensprache

    Landesdolmetscherzentrale für Gebärdensprache. https://landesdolmetscherzentrale-gebaerdensprache. de/lippenlesen/, 2020. [Accessed 16-09-2020]

    work page 2020

  7. [7]

    A survey of research on lipreading technology

    Mingfeng Hao, Mutallip Mamut, Nurbiya Yadikar, Alimjan Aysa, and Kurban Ubul. A survey of research on lipreading technology. IEEE Access, 8:204518–204544, 2020. 8

    work page 2020

  8. [8]

    Assael, Brendan Shillingford, Shimon Whiteson, and Nando de Freitas

    Yannis M. Assael, Brendan Shillingford, Shimon Whiteson, and Nando de Freitas. Lipnet: End-to-end sentence- level lipreading. arXiv:1611.01599, 2016

    arXiv 2016

Show all 38 references
  1. [9]

    Hoffman, Thomas Paine, Cían Hughes, Utsav Prabhu, Hank Liao, Hasim Sak, Kanishka Rao, Lorrayne Bennett, Marie Mulville, Ben Coppin, Ben Laurie, Andrew Senior, and Nando de Freitas

    Brendan Shillingford, Yannis Assael, Matthew W. Hoffman, Thomas Paine, Cían Hughes, Utsav Prabhu, Hank Liao, Hasim Sak, Kanishka Rao, Lorrayne Bennett, Marie Mulville, Ben Coppin, Ben Laurie, Andrew Senior, and Nando de Freitas. Large-scale visual speech recognition. arXiv:180...

    2018 arXiv

  2. [10]

    Hilder, Richard Harvey, and B

    S. Hilder, Richard Harvey, and B. Theobald. Comparison of human and machine-based lip-reading. 2009 Auditory-Visual Speech Processing (A VSP), pages 86–89, 01 2009

    2009

  3. [11]

    Nancy Tye-Murray, Sandra Hale, Brent Spehar, Joel Myerson, and Mitchell S. Sommers. Lipreading in school-age children: The roles of age, hearing status, and cognitive ability. Journal of Speech, Language, and Hearing Research, 57(2):556–565, April 2014

    2014

  4. [12]

    Deep Learning

    Ian Goodfellow, Yoshua Bengio, and Aaron Courville. Deep Learning . MIT Press, 2016. http://www. deeplearningbook.org

    2016

  5. [13]

    G. Cybenko. Approximation by superpositions of a sigmoidal function. Mathematics of Control, Signals, and Systems, 2(4):303–314, December 1989

    1989

  6. [14]

    Rectified-linear-unit- based deep learning for biomedical multi-label data

    Pu Wang, Ruiquan Ge, Xuan Xiao, Yunpeng Cai, Guoqing Wang, and Fengfeng Zhou. Rectified-linear-unit- based deep learning for biomedical multi-label data. Interdisciplinary Sciences: Computational Life Sciences , 9(3):419–422, November 2016

    2016

  7. [15]

    Pattern Recognition and Machine Learning (Information Science and Statistics)

    Christopher Bishop. Pattern Recognition and Machine Learning (Information Science and Statistics) . 10 2007

    2007

  8. [16]

    Lip reading using convolutional neural networks with and without pre-trained models

    Tayyip OZCAN and Alper BASTURK. Lip reading using convolutional neural networks with and without pre-trained models. Balkan Journal of Electrical and Computer Engineering , 7(2):195–201, April 2019

    2019

  9. [17]

    LeCun, B

    Y . LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, and L. D. Jackel. Backpropagation applied to handwritten zip code recognition. Neural Computation, 1(4):541–551, 1989

    1989

  10. [18]

    A 3d convolutional neural network for volumetric image semantic segmentation

    Hongya Lu, Haifeng Wang, Qianqian Zhang, Sang Won Yoon, and Daehan Won. A 3d convolutional neural network for volumetric image semantic segmentation. Procedia Manufacturing, 39:422–428, 2019

    2019

  11. [19]

    Lipreading with long short-term memory

    Michael Wand, Jan Koutník, and Jürgen Schmidhuber. Lipreading with long short-term memory. In 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , page 6115–6119. IEEE Press, 2016

    2016

  12. [20]

    Lip reading sentences in the wild

    Joon Son Chung, Andrew Senior, Oriol Vinyals, and Andrew Zisserman. Lip reading sentences in the wild. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages 3444–3453, 2017

    2017

  13. [21]

    Learning phrase representations using RNN encoder–decoder for statistical machine translation

    Kyunghyun Cho, Bart van Merriënboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. Learning phrase representations using RNN encoder–decoder for statistical machine translation. In Alessandro Moschitti, Bo Pang, and Walter Daelemans, edi...

    2014

  14. [22]

    Lrw-1000: A naturally-distributed large-scale benchmark for lip reading in the wild

    Shuang Yang, Yuanhang Zhang, Dalu Feng, Mingmin Yang, Chenhao Wang, Jingyun Xiao, Keyu Long, Shiguang Shan, and Xilin Chen. Lrw-1000: A naturally-distributed large-scale benchmark for lip reading in the wild. In 2019 14th IEEE International Conference on Automatic Face & Gestu...

    2019

  15. [23]

    3d convolutional neural networks for cross audio-visual matching recognition

    Amirsina Torfi, Seyed Mehdi Iranmanesh, Nasser Nasrabadi, and Jeremy Dawson. 3d convolutional neural networks for cross audio-visual matching recognition. IEEE Access, 5:22081–22091, 2017

    2017

  16. [24]

    Nickolay Shmyrev. V osk. https://github.com/alphacep/vosk-api, 2021. [Accessed 21-09-2021]

    2021

  17. [25]

    Gowda and Chun Yuan

    Shreyank N. Gowda and Chun Yuan. Colornet: Investigating the importance of color spaces for image classification. In C. V . Jawahar, Hongdong Li, Greg Mori, and Konrad Schindler, editors,ACCV (4), Lecture Notes in Computer Science, pages 581–596. Springer, 2018

    2018

  18. [26]

    Batch normalization: Accelerating deep network training by reducing internal covariate shift

    Sergey Ioffe and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv:1502.03167, 2015

    2015 arXiv

  19. [27]

    Dropout: a simple way to prevent neural networks from overfitting

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res., 15(1):1929–1958, January 2014

    1929

  20. [28]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In Yoshua Bengio and Yann LeCun, editors, 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, 2015. 9

    2015

  21. [29]

    youtube-dl

    Philipp Hagemeister Ricardo Garcia Gonzalez, Remita Amine. youtube-dl. https://github.com/ytdl-org/ youtube-dl, 2021. [Accessed 21-09-2021]

    2021

  22. [30]

    Converting video formats with ffmpeg

    Suramya Tomar. Converting video formats with ffmpeg. Linux J., 2006(146):10, June 2006

    2006

  23. [31]

    https://github.com/Zulko/moviepy, 2021

    Moviepy. https://github.com/Zulko/moviepy, 2021. [Accessed 21-09-2021]

    2021

  24. [32]

    The opencv library

    Gary Bradski. The opencv library. Dr . Dobb’s Journal, 25:122–125, 11 2000

    2000

  25. [33]

    Face recognition

    Adam Geitgey. Face recognition. https://github.com/ageitgey/face_recognition, 2021. [Accessed 21-09-2021]

    2021

  26. [34]

    Davis E. King. Dlib-ml: A machine learning toolkit. Journal of Machine Learning Research, 10(60):1755–1758, 2009

    2009

  27. [35]

    Adrian Rosebrock. Imutils. https://github.com/jrosebr1/imutils, 2021. [Accessed 21-09-2021]

    2021

  28. [36]

    Scikit-learn: Machine learning in python

    Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vanderplas, Alexandre Passos, David Cournapeau, Matthieu Brucher, Matthieu Perrot, and Édouard Duchesna...

    2011

  29. [37]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Köpf, Edward Yang, Zach DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fan...

    1912 arXiv

  30. [38]

    Lip reading sentences using deep learning with only visual cues

    Souheil Fenghour, Daqing Chen, Kun Guo, and Perry Xiao. Lip reading sentences using deep learning with only visual cues. IEEE Access, 8:215516–215530, 2020. 10

    2020

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.