Trainable Bilateral Filter Layer (PyTorch)

Last update: Dec 25, 2022

Overview

Trainable Bilateral Filter Layer (PyTorch)

This repository contains our GPU-accelerated trainable bilateral filter layer (three spatial and one range filter dimension) that can be directly included in any Pytorch graph, just as any conventional layer (FCL, CNN, ...). By calculating the analytical derivative of the bilateral filter with respect to its parameters and its input, the (so far) hyperparameters can be automatically optimized via backpropagation for a calculated loss.

Our corresponding paper Ultra low-parameter denoising: Trainable bilateral filter layers in computed tomography can be found on arXiv.

Implementation:

The general structure of the implementation follows the PyTorch documentation for creating custom C++ and CUDA extensions. The forward pass implementation of the layer is based on code from the Project MONAI framework, originally published under the Apache License, Version 2.0. The correct implementation of the analytical forward and backward pass can be verified by running the gradcheck.py script, comparing numerical gradients with the derived analytical gradient using the PyTorch built-in gradcheck function.

Setup:

The C++/CUDA implemented forward and backward functions are compiled via the setup.py script using setuptools:

Create and activate a python environment (python>=3.7).
Install Torch (tested versions: 1.7.1, 1.9.0).
Navigate into the extracted repo.
Compile/install the bilateral filter layer by calling

python setup.py install

Example scripts:

Try out the forward pass by running the example_filter.py (requires Matplotlib and scikit-image).
Run the gradcheck.py script to verify the correct gradient implementation.
Run example_optimization.py to optimize the parameters of a bilateral filter layer to automatically denoise an image.

Optimized bilateral filter prediction:

Citation:

If you find our code useful, please cite our work

@article{wagner2022ultra,
  title={Ultra low-parameter denoising: Trainable bilateral filter layers in computed tomography},
  author={Wagner, Fabian and Thies, Mareike and Gu, Mingxuan and Huang, Yixing and Pechmann, Sabrina and Patwari, Mayank and Ploner, Stefan and Aust, Oliver and Uderhardt, Stefan and Schett, Georg and Christiansen, Silke and Maier, Andreas},
  journal={arXiv preprint arXiv:2201.10345},
  year={2022}
}

You might also like...

📦 PyTorch based visualization package for generating layer-wise explanations for CNNs.

Implementation of the 😇 Attention layer from the paper, Scaling Local Self-Attention For Parameter Efficient Visual Backbones

HaloNet - Pytorch Implementation of the Attention layer from the paper, Scaling Local Self-Attention For Parameter Efficient Visual Backbones. This re

189 Nov 22, 2022

Comments

3D case example

Hi, I am grateful that you have shared such awesome codes. I have downloaded the codes and tested 2D case. It works well. But when I used 3D images, the results seemed weird. As shown below.

The input 3D image is like this:

Could you provide 3D demo case? Thank you very much!

opened by cs123951 2

Trainable Bilateral Filter Layer (PyTorch)

Related tags

Overview

Trainable Bilateral Filter Layer (PyTorch)

Implementation:

Setup:

Example scripts:

Optimized bilateral filter prediction:

Citation:

You might also like...

📦 PyTorch based visualization package for generating layer-wise explanations for CNNs.

PyTorch Implementation of NCSOFT's FastPitchFormant: Source-filter based Decomposed Modeling for Speech Synthesis

[CVPR 2021] Official PyTorch Implementation for "Iterative Filter Adaptive Network for Single Image Defocus Deblurring"

Pytorch Implementation of Continual Learning With Filter Atom Swapping (ICLR'22 Spolight) Paper

Predictive AI layer for existing databases.

Predictive AI layer for existing databases.

An abstraction layer for mathematical optimization solvers.

Understanding and Improving Encoder Layer Fusion in Sequence-to-Sequence Learning (ICLR 2021)

Implementation of the 😇 Attention layer from the paper, Scaling Local Self-Attention For Parameter Efficient Visual Backbones

Comments

3D case example

Releases(1.1.0)

1.1.0(Aug 1, 2022)

Owner

FabianWagner

A Python framework for developing parallelized Computational Fluid Dynamics software to solve the hyperbolic 2D Euler equations on distributed, multi-block structured grids.

Generate images from texts. In Russian. In PaddlePaddle

This project aim to create multi-label classification annotation tool to boost annotation speed and make it more easier.

Sequence lineage information extracted from RKI sequence data repo

Submodular Subset Selection for Active Domain Adaptation (ICCV 2021)

Tensorflow/Keras Plug-N-Play Deep Learning Models Compilation

Einshape: DSL-based reshaping library for JAX and other frameworks.

[NIPS 2021] UOTA: Improving Self-supervised Learning with Automated Unsupervised Outlier Arbitration.

The code for our paper Semi-Supervised Learning with Multi-Head Co-Training

OSLO: Open Source framework for Large-scale transformer Optimization

The 2nd place solution of 2021 google landmark retrieval on kaggle.

[ACL-IJCNLP 2021] Improving Named Entity Recognition by External Context Retrieving and Cooperative Learning

Official implementation of "Watermarking Images in Self-Supervised Latent-Spaces"

Web mining module for Python, with tools for scraping, natural language processing, machine learning, network analysis and visualization.

Improving XGBoost survival analysis with embeddings and debiased estimators

A set of tools for creating and testing machine learning features, with a scikit-learn compatible API

Uni-Fold: Training your own deep protein-folding models.

Hard cater examples from Hopper ICLR paper

StackRec: Efficient Training of Very Deep Sequential Recommender Models by Iterative Stacking

source code of “Visual Saliency Transformer” (ICCV2021)