edufonseca / shift_sec

Code for the paper "Improving Sound Event Classification by Increasing Shift Invariance in Convolutional Neural Networks".

Improving Sound Event Classification by
Increasing Shift Invariance in Convolutional Neural Networks

This repository contains the TensorFlow implementation of the models and pooling mechanisms proposed in the following paper. If you use this code or part of it, please cite:

Eduardo Fonseca, Andres Ferraro, Xavier Serra, "Improving Sound Event Classification by Increasing Shift Invariance in Convolutional Neural Networks", arXiv:2107.00623, 2021.

Specifically, the implementations included are

• models: VGG41 and VGG42 (see model.py)
• pooling mechanisms to increase shift invariance in convolutional neural networks:
  • Trainable Low-Pass Filters (TLPF) and BlurPool, inspired by [1], (see lpf.py)
  • Adaptive Polyphase Sampling (APS), inspired by [2], (see aps.py)

Citation

@article{fonseca2021shift,
  title={Improving Sound Event Classification by Increasing Shift Invariance in Convolutional Neural Networks},
  author={Fonseca, Eduardo and Ferraro, Andres and Serra, Xavier},
  journal={arXiv preprint arXiv:2107.00623},
  year={2021}
}

Questions?

Please contact Edu Fonseca with comments and questions efonseca at google dot com.

References

[1] R. Zhang, “Making convolutional networks shift-invariant again,” in International Conference on Machine Learning. PMLR, 2019, pp. 7324–7334.

[2] A. Chaman and I. Dokmanic, “Truly shift-invariant convolutional neural networks,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 3773–3783.