Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

(Un)paired signal-to-signal translation with 1D conditional GANs

Eric Easthope

Abstract

I show that a one-dimensional (1D) conditional generative adversarial network (cGAN) with an adversarial training architecture is capable of unpaired signal-to-signal ("sig2sig") translation. Using a simplified CycleGAN model with 1D layers and wider convolutional kernels, mirroring WaveGAN to reframe two-dimensional (2D) image generation as 1D audio generation, I show that recasting the 2D image-to-image translation task to a 1D signal-to-signal translation task with deep convolutional GANs is possible without substantial modification to the conventional U-Net model and adversarial architecture developed as CycleGAN. With this I show for a small tunable dataset that noisy test signals unseen by the 1D CycleGAN model and without paired training transform from the source domain to signals similar to paired test signals in the translated domain, especially in terms of frequency, and I quantify these differences in terms of correlation and error.

(Un)paired signal-to-signal translation with 1D conditional GANs

Abstract

I show that a one-dimensional (1D) conditional generative adversarial network (cGAN) with an adversarial training architecture is capable of unpaired signal-to-signal ("sig2sig") translation. Using a simplified CycleGAN model with 1D layers and wider convolutional kernels, mirroring WaveGAN to reframe two-dimensional (2D) image generation as 1D audio generation, I show that recasting the 2D image-to-image translation task to a 1D signal-to-signal translation task with deep convolutional GANs is possible without substantial modification to the conventional U-Net model and adversarial architecture developed as CycleGAN. With this I show for a small tunable dataset that noisy test signals unseen by the 1D CycleGAN model and without paired training transform from the source domain to signals similar to paired test signals in the translated domain, especially in terms of frequency, and I quantify these differences in terms of correlation and error.
Paper Structure (9 sections, 4 figures)

This paper contains 9 sections, 4 figures.

Table of Contents

  1. Introduction
  2. Background.
  3. Objective.
  4. Methods & Materials
  5. Model.
  6. Dataset.
  7. Results
  8. Discussion
  9. Conclusion

Figures (4)

  • Figure 1: Sixteen randomly sampled dataset elements used for 1D CycleGAN training between one-dimensional signal domains (blue/purple lines) with arbitrary scales.
  • Figure 2: Paired signal-to-signal translations by the simplified 1D CycleGAN architecture against a small unseen four-element test dataset in the time domain ($r$-values in discussion).
  • Figure 3: Paired signal-to-signal translations by the simplified 1D CycleGAN architecture against a small unseen four-element test dataset in the frequency domain after discrete Fast Fourier Transform. Frequency-wise signals seem to match more than time-wise signals above ($r$-values in discussion).
  • Figure 4: Signal-to-signal translations (supplementary anonymized dataset) by the simplified CycleGAN architecture, which combines elements of the pix2pix architecture with a different training pattern and loss for unpaired signal data. Differences in reconstruction are shown at both signal $X$ and signal $Y$ translation domains (compared as solid/transparent lines).