Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
162 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Blind Source Separation of Single-Channel Mixtures via Multi-Encoder Autoencoders (2309.07138v3)

Published 31 Aug 2023 in eess.SP and cs.LG

Abstract: The task of blind source separation (BSS) involves separating sources from a mixture without prior knowledge of the sources or the mixing system. Single-channel mixtures and non-linear mixtures are a particularly challenging problem in BSS. In this paper, we propose a novel method for addressing BSS with single-channel non-linear mixtures by leveraging the natural feature subspace specialization ability of multi-encoder autoencoders. During the training phase, our method unmixes the input into the separate encoding spaces of the multi-encoder network and then remixes these representations within the decoder for a reconstruction of the input. Then to perform source inference, we introduce a novel encoding masking technique whereby masking out all but one of the encodings enables the decoder to estimate a source signal. To this end, we also introduce a sparse mixing loss that encourages sparse remixing of source encodings throughout the decoder and a so-called zero reconstruction loss on the decoder for coherent source estimations. To analyze and evaluate our method, we conduct experiments on a toy dataset, designed to demonstrate this property of feature subspace specialization, and with real-world biosignal recordings from a polysomnography sleep study for extracting respiration from electrocardiogram and photoplethysmography signals.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (106)
  1. Blind image separation using the jade method. Engineering Proceedings 14, 1 (2022).
  2. Almeida, L. Misep - linear and nonlinear ica based on mutual information. Journal of Machine Learning Research 4 (01 2003).
  3. A blind source separation method for chemical sensor arrays based on a second order mixing model. 2015 23rd European Signal Processing Conference (EUSIPCO) (2015), 933–937.
  4. A survey of artificial intelligence approaches in blind source separation. Neurocomput. 561, C (01 2024).
  5. An end-to-end and accurate ppg-based respiratory rate estimation approach using cycle generative adversarial networks. In 2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) (2021), pp. 744–747.
  6. Removing artifacts from electrocardiographic signals using independent components analysis. Neurocomputing 22, 1 (1998), 173–186.
  7. An information-maximization approach to blind separation and blind deconvolution. Neural computation 7 (12 1995), 1129–59.
  8. Learning independent features with adversarial nets for non-linear ica. arXiv: Machine Learning (2017).
  9. Cardoso, J.-F. Multidimensional independent component analysis. Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP ’98 (Cat. No.98CH36181) 4 (1998), 1941–1944 vol.4.
  10. An assessment of algorithms to estimate respiratory rate from the electrocardiogram and photoplethysmogram. Physiological Measurement 37, 4 (03 2016), 610.
  11. Dual-Path Transformer Network: Direct Context-Aware Modeling for End-to-End Monaural Speech Separation. In Proc. Interspeech 2020 (2020), pp. 2642–2646.
  12. Racial/ethnic differences in sleep disturbances: The multi-ethnic study of atherosclerosis (mesa). Sleep 38, 6 (06 2015), 877–888.
  13. Deep attractor network for single-microphone speaker separation. 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2016), 246–250.
  14. New algorithms for non-negative matrix factorization in applications to blind source separation. In 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings (2006), vol. 5, pp. V–V.
  15. Handbook of Blind Source Separation, Independent Component Analysis and Applications. Academic Press, 02 2010.
  16. Rapid extraction of respiratory waveforms from photoplethysmography: A deep corr-encoder approach. Biomedical Signal Processing and Control 85 (2023), 104992.
  17. Deville, Y. Blind source separation and blind mixture identification methods. In Wiley Encyclopedia of Electrical and Electronics Engineering. John Wiley & Sons, Inc., Hoboken, NJ, USA, 2016, pp. 1–33.
  18. Deville, Y. From separability/identifiability properties of bilinear and linear-quadratic mixture matrix factorization to factorization algorithms. Digital Signal Processing 87 (2019), 21–33.
  19. An overview of blind source separation methods for linear-quadratic and post-nonlinear mixtures. vol. 9237, pp. 155–167.
  20. Blind signal separation and identification of mixtures of images. In 2009 Conference Record of the Forty-Third Asilomar Conference on Signals, Systems and Computers (2009), pp. 337–342.
  21. Source separation in chemical analysis : Recent achievements and perspectives. Signal Processing Magazine, IEEE 31 (05 2014), 135–146.
  22. Blind source separation of overdetermined linear-quadratic mixtures. pp. 263–270.
  23. Blind source separation in nonlinear mixtures: Separability and a basic algorithm. IEEE Transactions on Signal Processing PP (05 2017), 1–1.
  24. Protocaps: A fast and non-iterative capsule network routing method. Transactions on Machine Learning Research (2023).
  25. An online subspace denoising algorithm for maternal ecg removal from fetal ecg signals. Iranian Journal of Science and Technology, Transactions of Electrical Engineering 41 (04 2017).
  26. "double-dip": Unsupervised image decomposition via coupled deep-image-priors. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (06 2019).
  27. Kernel feature spaces and nonlinear blind souce separation. In Advances in Neural Information Processing Systems (2001), T. Dietterich, S. Becker, and Z. Ghahramani, Eds., vol. 14, MIT Press.
  28. Deep clustering: Discriminative embeddings for segmentation and separation. In 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2016), pp. 31–35.
  29. Reducing the dimensionality of data with neural networks. Science 313, 5786 (2006), 504–507.
  30. Respiratory sinus arrhythmia in humans: how breathing pattern modulates heart rate. The American journal of physiology 241 4 (1981), H620–9.
  31. Hoshen, Y. Towards unsupervised single-channel blind source separation using adversarial pair unmix-and-remix. In ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2019), pp. 3272–3276.
  32. Hyvarinen, A. Fast and robust fixed-point algorithms for independent component analysis. IEEE Transactions on Neural Networks 10, 3 (1999), 626–634.
  33. Hyvärinen, A. Independent component analysis: recent advances. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 371, 1984 (2013), 20110534.
  34. One-unit learning rules for independent component analysis. In Advances in Neural Information Processing Systems (1996), M. Mozer, M. Jordan, and T. Petsche, Eds., vol. 9, MIT Press.
  35. Independent component analysis: algorithms and applications. Neural Networks 13, 4 (2000), 411–430.
  36. Nonlinear independent component analysis: Existence and uniqueness results. Neural Networks 12, 3 (1999), 429–439.
  37. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In Proceedings of the 32nd International Conference on Machine Learning (Lille, France, 07 2015), F. Bach and D. Blei, Eds., vol. 37 of Proceedings of Machine Learning Research, PMLR, pp. 448–456.
  38. Nonlinear cardio-respiratory interactions revealed by time-phase bispectral analysis. Physics in medicine and biology 49 18 (2004), 4407–25.
  39. Fetal ecg extraction from time-varying and low-rank noninvasive maternal abdominal recordings. Physiological Measurement 39 (11 2018).
  40. Source separation with deep generative priors. In Proceedings of the 37th International Conference on Machine Learning (07 2020), H. D. III and A. Singh, Eds., vol. 119 of Proceedings of Machine Learning Research, PMLR, pp. 4724–4735.
  41. Structured sparse principal component analysis. In Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics (Chia Laguna Resort, Sardinia, Italy, 13–15 05 2010), Y. W. Teh and M. Titterington, Eds., vol. 9 of Proceedings of Machine Learning Research, PMLR, pp. 366–373.
  42. Supervised determined source separation with multichannel variational autoencoder. Neural Computation 31, 9 (09 2019), 1891–1914.
  43. Influence of forced respiration on nonlinear dynamics in heart rate variability. Am J Physiol 272, 4 Pt 2 (04 1997), R1149–54.
  44. Kofidis, E. Blind source separation: Fundamentals and recent advances (a tutorial overview presented at sbrt-2001), 2016.
  45. Single-channel signal separation and deconvolution with generative adversarial networks. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (Aug. 2019), IJCAI-2019, International Joint Conferences on Artificial Intelligence Organization.
  46. Imagenet classification with deep convolutional neural networks. Communications of the ACM 60, 6 (2012), 84–90.
  47. Kuleshov, V. Fast algorithms for sparse principal component analysis based on rayleigh quotient iteration. In Proceedings of the 30th International Conference on Machine Learning (Atlanta, Georgia, USA, 17–19 Jun 2013), S. Dasgupta and D. McAllester, Eds., vol. 28 of Proceedings of Machine Learning Research, PMLR, pp. 1418–1425.
  48. Fetal electrocardiogram extraction by source subspace separation.
  49. Learning the parts of objects by non-negative matrix factorization. Nature 401, 6755 (1999), 788–791.
  50. End-to-end deep learning architecture for separating maternal and fetal ecgs using w-net. IEEE Access 10 (2022), 39782–39788.
  51. Estimation of heart rate and respiratory rate from ppg signal using complementary ensemble empirical mode decomposition with both independent component analysis and non-negative matrix factorization. Sensors 20, 11 (2020).
  52. Intelligent blind source separation technology based on otfs modulation for leo satellite communication. China Communications 19, 7 (2022), 89–99.
  53. Design and implementation of blind source separation based on bp neural network in space-based ais. Frontiers in Space Technologies 2 (2021).
  54. Blind source separation in noisy environment and applications in satellite communication anti-jamming. In 2009 Asia-Pacific Conference on Computational Intelligence and Industrial Applications (PACIIA) (2009), vol. 2, pp. 96–99.
  55. Monitoring of respiratory and heart rates using a fibre-optic sensor. Medical and Biological Engineering and Computing 30 (1992), 533–537.
  56. Heart rate and respiratory rhythm dynamics on ascent to high altitude. British Heart Journal 74 (1995), 390–396.
  57. Separate in latent space: Unsupervised single image layer separation. In Proceedings of the AAAI Conference on Artificial Intelligence (2020), vol. 34, pp. 11661–11668.
  58. Dual-path rnn: Efficient long sequence modeling for time-domain single-channel speech separation. In ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2020), pp. 46–50.
  59. Tasnet: Time-domain audio separation network for real-time, single-channel speech separation. 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2017), 696–700.
  60. Conv-tasnet: Surpassing ideal time-frequency magnitude masking for speech separation. IEEE/ACM Transactions on Audio, Speech, and Language Processing PP (05 2019), 1–1.
  61. NeuroKit2: A python toolbox for neurophysiological signal processing. Behavior Research Methods 53, 4 (02 2021), 1689–1696.
  62. Unsupervised neural decoding of signals recorded by thin-film electrode arrays implanted in muscles using autoencoding with a physiologically derived optimisation criterion. Biomedical Signal Processing and Control 86 (2023), 105178.
  63. Breath rate variability (brv) - a novel measure to study the meditation effects. International Journal of Yoga Accepted (01 2017).
  64. Unsupervised blind source separation with variational auto-encoders. In 2021 29th European Signal Processing Conference (EUSIPCO) (2021), pp. 311–315.
  65. Cass: Cross adversarial source separation via autoencoder. ArXiv abs/1905.09877 (2019).
  66. Information-theoretic analysis of cardio-respiratory interactions in heart failure patients: Effects of arrhythmias and cardiac resynchronization therapy. Entropy 25 (07 2023).
  67. Blind source separation (bss) of mixed maternal and fetal electrocardiogram (ecg) signal: A comparative study. Procedia Computer Science 176 (2020), 582–591. Knowledge-Based and Intelligent Information & Engineering Systems: Proceedings of the 24th International Conference KES2020.
  68. Respnet: A deep learning model for extraction of respiration from photoplethysmogram. In 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (2019), pp. 5556–5559.
  69. Using sparse autoencoders to perform blind source separation of high-density myoelectric signal. In 2023 IEEE 33rd International Workshop on Machine Learning for Signal Processing (MLSP) (2023), pp. 1–6.
  70. Hybrid transformers for music source separation, 2022.
  71. Extraction of respiration signal from ecg for respiratory rate estimation. IET Conference Proceedings (01 2015), 58 (5 .)–58 (5 .)(1).
  72. Weight normalization: A simple reparameterization to accelerate training of deep neural networks, 2016.
  73. A review of fetal ecg signal processing issues and promising directions. The open pacing, electrophysiology & therapy journal 3 (01 2010), 4–20.
  74. What ica provides for ecg processing: Application to noninvasive fetal ecg extraction. In 2006 IEEE International Symposium on Signal Processing and Information Technology (2006), pp. 656–661.
  75. Meta-learning extractors for music source separation. ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2020), 816–820.
  76. Kernel principal component analysis. In Artificial Neural Networks — ICANN’97 (Berlin, Heidelberg, 1997), W. Gerstner, A. Germond, M. Hasler, and J.-D. Nicoud, Eds., Springer Berlin Heidelberg, pp. 583–588.
  77. Fetal electrocardiogram extraction using dual-path source separation of single-channel non-invasive abdominal recordings. IEEE Transactions on Biomedical Engineering 70, 1 (2023), 283–295.
  78. Independent component analysis and blind signal separation: Theory, algorithms and applications. Learning and Nonlinear Models 10 (01 2012), 51–69.
  79. An overview of variational autoencoders for source separation, finance, and bio-signal applications. Entropy 24, 1 (2022).
  80. Spline neural networks for blind separation of post-nonlinear-linear mixtures. Circuits and Systems I: Regular Papers, IEEE Transactions on 51 (05 2004), 817–829.
  81. Sparse nonnegative matrix factorization applied to microarray data sets. In Independent Component Analysis and Blind Signal Separation (Berlin, Heidelberg, 2006), J. Rosca, D. Erdogmus, J. C. Príncipe, and S. Haykin, Eds., Springer Berlin Heidelberg, pp. 254–261.
  82. Adversarial semi-supervised audio source separation applied to singing voice extraction. In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2018), IEEE, pp. 2391–2395.
  83. Wave-u-net: A multi-scale neural network for end-to-end audio source separation. In Proceedings of the 19th International Society for Music Information Retrieval Conference, ISMIR 2018, Paris, France, 2018 (09 2018), E. Gómez, X. Hu, E. Humphrey, and E. Benetos, Eds., pp. 334–340.
  84. Attention is all you need in speech separation. In ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2021), pp. 21–25.
  85. Single-channel blind source separation based on attentional generative adversarial network. Journal of Ambient Intelligence and Humanized Computing 13 (03 2022), 1–8.
  86. Chou’s electrocardiography in clinical practice: adult and pediatric. Elsevier Health Sciences, 2008.
  87. Nonlinear blind source separation using a radial basis function network.
  88. Why self-attention? a targeted evaluation of neural machine translation architectures. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (Brussels, Belgium, 10 2018), Association for Computational Linguistics, pp. 4263–4272.
  89. A multi-encoder variational autoencoder controls multiple transformational features in single-cell image analysis. Commun Biol 5 (2022), 255.
  90. Umesh, P. Image processing in python. CSI Communications 23 (2012).
  91. Heartpy: A novel heart rate algorithm for the analysis of noisy signals. Transportation Research Part F: Traffic Psychology and Behaviour 66 (2019), 368–378.
  92. A comparative study of ecg-derived respiration in ambulatory monitoring using the single-lead ecg. Scientific Reports 10 (03 2020), 5704.
  93. Attention is all you need. In Proceedings of the 31st International Conference on Neural Information Processing Systems (2017), pp. 6000–6010.
  94. Performance measurement in blind audio source separation. IEEE Transactions on Audio, Speech, and Language Processing 14, 4 (2006), 1462–1469.
  95. Context-aware neural machine translation learns anaphora resolution. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) (Melbourne, Australia, 07 2018), Association for Computational Linguistics, pp. 1264–1274.
  96. Analyzing multi-head self-attention: Specialized heads do the heavy lifting, the rest can be pruned. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (Florence, Italy, 07 2019), Association for Computational Linguistics, pp. 5797–5808.
  97. Image Quality Assessment: From Error Visibility to Structural Similarity. IEEE Transactions on Image Processing 13, 4 (04 2004), 600–612.
  98. Application of kernel principal component analysis for single-lead-ecg-derived respiration. IEEE Transactions on Biomedical Engineering 59, 4 (2012), 1169–1176.
  99. General multilayer perceptron demixer scheme for nonlinear blind signal separation. Vision, Image and Signal Processing, IEE Proceedings - 149 (11 2002), 253–262.
  100. Group normalization. In Proceedings of the European Conference on Computer Vision (ECCV) (09 2018).
  101. A new blind image source separation algorithm based on feedback sparse component analysis. Signal Processing 93, 1 (2013), 288–296.
  102. Blind source separation for satellite communication anti-jamming. Wireless and Satellite Systems 280 (2019), 717–726.
  103. The national sleep research resource: Towards a sleep data commons. Journal of the American Medical Informatics Association 25, 10 (10 2018), 1351–1358.
  104. Mossformer: Pushing the performance limit of monaural speech separation using gated single-head transformer with convolution-augmented joint self-attentions. CoRR abs/2302.11824 (2023).
  105. The key technology of blind source separation of satellite-based ais. Procedia Engineering 29 (2012), 3737–3741. 2012 International Workshop on Information and Electronics Engineering.
  106. Deep adversarial decomposition: A unified framework for separating superimposed images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (06 2020).

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now