[IEEE Transactions on Computational Imaging] Self-Gated Memory Recurrent Network for Efficient Scalable HDR Deghosting

Last update: Dec 29, 2021

Related tags

Overview

Few-shot Deep HDR Deghosting

This repository contains code and pretrained models for our paper: Self-Gated Memory Recurrent Network for Efficient Scalable HDR Deghosting accepted to IEEE Transactions on Computational Imaging.

It has been tested on RTX 6000 with Tensorflow 2.3.4.

Installation and Setup

Docker Environment:

Getting base image:

$ docker pull tensorflow/tensorflow:2.3.4-gpu

Running base image:

$ docker run --rm -it tensorflow/tensorflow:2.3.4-gpu bash

Installing dependencies:

(docker)# apt update
(docker)# apt install -y ffmpeg libsm6 libxext6 libxrender-dev
(docker)# pip install opencv-python

Datasets:

The Kalantari dataset (SIG17) can be downloaded here, and the Prabhakar dataset (ICCP19) can be downloaded here.

Download the required dataset and extract it in the dataset folder.

Training

To view all training options, run

$ python main.py --help

To train a Bidirectional SGM model with the default configuration, run

$ python main.py --rtx

Inference

To evaluate the pretrained model on the Kalantari17 dataset, run

$ python val.py --rtx --weights pretrained_weights/UCSD/bidirectional.sgm/bidirectional.sgm.tf

Citation

When citing this work, you should use the following Bibtex:

@ARTICLE{9540317, 
    author={Kathirvel, Ram Prabhakar and Agrawal, Susmit and Radhakrishnan, Venkatesh Babu},
    journal={IEEE Transactions on Computational Imaging}, 
    title={Self-Gated Memory Recurrent Network for Efficient Scalable HDR Deghosting}, 
    year={2021},
    volume={},
    number={},
    pages={1-1},
    doi={10.1109/TCI.2021.3112920}
}

PyTorch Code of "Memory In Memory: A Predictive Neural Network for Learning Higher-Order Non-Stationarity from Spatiotemporal Dynamics"

Memory In Memory Networks It is based on the paper Memory In Memory: A Predictive Neural Network for Learning Higher-Order Non-Stationarity from Spati

12 May 30, 2022

[CVPR 21] Vectorization and Rasterization: Self-Supervised Learning for Sketch and Handwriting, IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2021.

Vectorization and Rasterization: Self-Supervised Learning for Sketch and Handwriting, CVPR 2021. Ayan Kumar Bhunia, Pinaki nath Chowdhury, Yongxin Yan

44 Dec 12, 2022

The official implementation of the IEEE S&P`22 paper "SoK: How Robust is Deep Neural Network Image Classification Watermarking".

Watermark-Robustness-Toolbox - Official PyTorch Implementation This repository contains the official PyTorch implementation of the following paper to

49 Dec 19, 2022

Comments

Weights for the constructed scalable dataset

hi @KRamPrabhakar @Susmit-A

Thanks for the excellent work. You have provided the weights of HDRRNN on the UCSD dataset. Could you also share the weights of the HDRRNN on the scalable HDR dataset (https://val.cds.iisc.ac.in/HDR/HDRD/)?

Best, Hanwei

opened by h4nwei 0

[IEEE Transactions on Computational Imaging] Self-Gated Memory Recurrent Network for Efficient Scalable HDR Deghosting

Related tags

Overview

Few-shot Deep HDR Deghosting

Installation and Setup

Docker Environment:

Datasets:

Training

Inference

Citation

You might also like...

PyTorch Code of "Memory In Memory: A Predictive Neural Network for Learning Higher-Order Non-Stationarity from Spatiotemporal Dynamics"

[CVPR 21] Vectorization and Rasterization: Self-Supervised Learning for Sketch and Handwriting, IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2021.

Implementation of "Efficient Regional Memory Network for Video Object Segmentation" (Xie et al., CVPR 2021).

Pytorch Code for "Medical Transformer: Gated Axial-Attention for Medical Image Segmentation"

A PyTorch Implementation of Gated Graph Sequence Neural Networks (GGNN)

A PyTorch Implementation of Gated Graph Sequence Neural Networks (GGNN)

Episodic-memory - Ego4D Episodic Memory Benchmark

(IEEE TIP 2021) Regularized Densely-connected Pyramid Network for Salient Instance Segmentation

The official implementation of the IEEE S&P`22 paper "SoK: How Robust is Deep Neural Network Image Classification Watermarking".

Comments

Weights for the constructed scalable dataset

Owner

Susmit Agrawal

PyTorch implementation of Deep HDR Imaging via A Non-Local Network (TIP 2020).

Segcache: a memory-efficient and scalable in-memory key-value cache for small objects

Pytorch implementation of various High Dynamic Range (HDR) Imaging algorithms

Implementation of a memory efficient multi-head attention as proposed in the paper, "Self-attention Does Not Need O(n²) Memory"

The Dual Memory is build from a simple CNN for the deep memory and Linear Regression fro the fast Memory

Official implementation of NLOS-OT: Passive Non-Line-of-Sight Imaging Using Optimal Transport (IEEE TIP, accepted)

Keyhole Imaging: Non-Line-of-Sight Imaging and Tracking of Moving Objects Along a Single Optical Path

Space Time Recurrent Memory Network - Pytorch

HDR Video Reconstruction: A Coarse-to-fine Network and A Real-world Benchmark Dataset (ICCV 2021)

Implementation of Bidirectional Recurrent Independent Mechanisms (Learning to Combine Top-Down and Bottom-Up Signals in Recurrent Neural Networks with Attention over Modules)