A Fast, Scalable, and Reliable Deghosting Method for Extreme Exposure Fusion

@article{Prabhakar2019AFS,
  title={A Fast, Scalable, and Reliable Deghosting Method for Extreme Exposure Fusion},
  author={K. Prabhakar and Rajat Arora and Adhitya Swaminathan and Kunal Pratap Singh and R. Venkatesh Babu},
  journal={2019 IEEE International Conference on Computational Photography (ICCP)},
  year={2019},
  pages={1-8}
}
HDR fusion of extreme exposure images with complex camera and object motion is a challenging task. Existing patch-based optimization techniques generate noisy and/or blurry results with undesirable artifacts for difficult scenarios. Additionally, they are computationally intensive and have high execution times. Recently proposed CNN-based methods offer fast alternatives, but still fail to generate artifact-free results for extreme exposure images. Furthermore, they do not scale to an arbitrary… 
Towards Practical and Efficient High-Resolution HDR Deghosting with CNN
TLDR
This paper presents a deep neural network based approach to generate high-quality ghost-free HDR for high-resolution images and shows that the proposed method achieves state-of-the-art performance in terms of accuracy and speed.
Image Adjustment for Multi-Exposure Images Based on Convolutional Neural Networks
TLDR
An image adjustment method for multi-exposure images based on convolutional neural networks (CNNs) that outperforms state-of-the-art fusion methods in the peak signal-to-noise ratio and is applied for various fusion methods as pre-processing, and results show obviously reducing artifacts.
Robust High Dynamic Range (HDR) Imaging with Complex Motion and Parallax
TLDR
A Pyramidal Alignment and Masked merging network (PAMnet) that learns to synthesize HDR images from input low dynamic range (LDR) images in an end-to-end manner and can produce ghosting-free HDR results in the presence of large disparity and motion is proposed.
Single-Shot High Dynamic Range Imaging via Multiscale Convolutional Neural Network
TLDR
To effectively estimate the high-quality HDR images, a robust loss function is developed that considers the human visual system (HVS), color perception model, and multiscale contrast and can achieve synthesis results of significantly higher quality than conventional algorithms.
Intensity Mapping Functions For HDR Panorama Imaging: Weighted Histogram Averaging
TLDR
Extensive experiments indicate that the proposed WHA algorithm significantly surpasses the related state-of-the-art intensity mapping methods and preserves details in the brightest and darkest regions of the input images well.
Color Mapping Functions For HDR Panorama Imaging: Weighted Histogram Averaging
TLDR
A novel color mapping algorithm is first proposed by introducing a new concept of weighted histogram averaging (WHA) which significantly surpasses the related state-of-the-art color mapping methods.
SJ-HD^2R: Selective Joint High Dynamic Range and Denoising Imaging for Dynamic Scenes
TLDR
A joint HDR and denoising pipeline, containing a pre-denoising network (PreDNNet) to adaptively denoise input LDRs by exploiting exposure priors and a pyramid cascading fusion network (PCFNet), introducing an attention mechanism and cascading structure in a multi-scale manner.
Two-Exposure Image Fusion Based on Optimized Adaptive Gamma Correction
TLDR
This work generates a more helpful intermediate virtual image for fusion using the proposed Optimized Adaptive Gamma Correction (OAGC) to have better contrast, saturation, and well-exposedness.
MIEHDR CNN: Main Image Enhancement based Ghost-Free High Dynamic Range Imaging using Dual-Lens Systems
TLDR
The MIEHDR CNN model, which consists of three subnets, is implemented by MindSpore and experimental results show that it can outperform related methods largely and avoid ghost artifacts.
FlexHDR: Modelling Alignment and Exposure Uncertainties for Flexible HDR Imaging
TLDR
This paper presents a new HDR imaging technique that specifically models alignment and exposure uncertainties to produce high quality HDR results and introduces a progressive, multi-stage image fusion approach that can flexibly merge any number of LDR images in a permutation-invariant manner.
...
...

References

SHOWING 1-10 OF 24 REFERENCES
Deep high dynamic range imaging of dynamic scenes
TLDR
A convolutional neural network is used as the learning model and three different system architectures are compared to model the HDR merge process to demonstrate the performance of the system by producing high-quality HDR images from a set of three LDR images.
DeepFuse: A Deep Unsupervised Approach for Exposure Fusion with Extreme Exposure Image Pairs
TLDR
An unsupervised deep learning framework for MEF utilizing a no-reference quality metric as loss function utilizing a novel CNN architecture trained to learn the fusion operation without reference ground truth image is proposed.
Deep High Dynamic Range Imaging with Large Foreground Motions
TLDR
This paper proposes the first non-flow-based deep framework for high dynamic range (HDR) imaging of dynamic scenes with large-scale foreground motions, and produces excellent results where color artifacts and geometric distortions are significantly reduced compared to existing state-of-the-art methods.
Robust Multi-Exposure Image Fusion: A Structural Patch Decomposition Approach
TLDR
The proposed algorithm not only outperforms previous MEF algorithms on static scenes but also consistently produces high quality fused images with little ghosting artifacts for dynamic scenes and maintains a lower computational cost compared with the state-of-the-art deghosting schemes.
Freehand HDR Imaging of Moving Scenes with Simultaneous Resolution Enhancement
TLDR
The alignment step is performed by a modern energy‐based optic flow approach that takes into account the varying exposure conditions, and produces dense displacement fields with subpixel precision, which can handle arbitrary complex motion patterns, caused by severe camera shake and moving objects.
The State of the Art in HDR Deghosting: A Survey and Evaluation
TLDR
A taxonomy of deghosting algorithms is proposed which can be used to group existing and future algorithms into meaningful classes, and the results of a subjective experiment are shared which aims to evaluate various state‐of‐the‐art de ghosting algorithms.
Robust patch-based hdr reconstruction of dynamic scenes
TLDR
This paper proposes a new approach to HDR reconstruction that draws information from all the exposures but is more robust to camera/scene motion than previous techniques and presents results that show considerable improvement over previous approaches.
Overview of State-of-the-Art Algorithms for Stack-Based High-Dynamic Range (HDR) Imaging
TLDR
An overview of state-of-the-art deghosting algorithms for stackbased HDR imaging is presented and some of the tradeoffs of each are discussed.
Artifact-free High Dynamic Range imaging
TLDR
This work presents a technique capable of dealing with a large amount of movement in the scene: it finds, in all the available exposures, patches consistent with a reference image previously selected from the stack and generates the HDR image by averaging the radiance estimates of all such regions.
Exposure Fusion
TLDR
This work proposes a technique for fusing a bracketed exposure sequence into a high quality image, without converting to HDR first, which avoids camera response curve calibration and is computationally efficient.
...
...