HDNet: High-resolution Dual-domain Learning for Spectral Compressive Imaging (2022-03-04T00:00:00.000000Z)

TL;DR

A high-resolution dual-domain learning network (HDNet) for HSI reconstruction with HR pixel-level attention and frequency-level refinement that achieves SOTA performance on simulated and real HSI datasets.

Abstract

The rapid development of deep learning provides a better solution for the end-to-end reconstruction of hyperspectral image (HSI). However, existing learning-based methods have two major defects. Firstly, networks with self-attention usually sacrifice internal resolution to balance model performance against complexity, losing fine-grained high-resolution (HR) features. Secondly, even if the optimization focusing on spatial-spectral domain learning (SDL) converges to the ideal solution, there is still a significant visual difference between the reconstructed HSI and the truth. So we propose a high-resolution dual-domain learning network (HDNet) for HSI reconstruction. On the one hand, the proposed HR spatial-spectral attention module with its efficient feature fusion provides continuous and fine pixel-level features. On the other hand, frequency domain learning (FDL) is introduced for HSI reconstruction to narrow the frequency domain discrepancy. Dynamic FDL supervision forces the model to reconstruct fine-grained frequencies and compensate for excessive smoothing and distortion caused by pixel-level losses. The HR pixel-level attention and frequency-level refinement in our HDNet mutually promote HSI perceptual quality. Extensive quantitative and qualitative experiments show that our method achieves SOTA performance on simulated and real HSI datasets. https://github.com/Huxiaowan/HDNet

Authors

L. Gool

68 papers

Radu Timofte

49 papers

Haoqian Wang

8 papers

TL;DR

Abstract

Authors

References58 items

Learning to Generate Realistic Noisy Images via Pixel-level Noise-aware Adversarial Training

Flow-Guided Sparse Transformer for Video Deblurring

RFormer: Transformer-Based Generative Adversarial Network for Real Fundus Image Restoration on a New Clinical Benchmark

Mask-guided Spectral-wise Transformer for Efficient Hyperspectral Image Reconstruction

Polarized Self-Attention: Towards High-quality Pixel-wise Regression

Deep Gaussian Scale Mixture Prior for Spectral Compressive Imaging

Focal Frequency Loss for Image Reconstruction and Synthesis

GAP-net for Snapshot Compressive Imaging

FakeRetouch: Evading DeepFakes Detection via the Guidance of Deliberate Noise

End-to-End Low Cost Compressive Spectral Imaging with Spatial-Spectral Self-Attention

Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains

DNU: Deep Non-Local Unrolling for Computational Spectral Imaging

Unsupervised Real-world Image Super Resolution via Domain-distance Aware Training

Learning Delicate Local Representations for Multi-Person Pose Estimation

CNN-Generated Images Are Surprisingly Easy to Spot… for Now

Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation

Frequency Separation for Real-World Super-Resolution

Computational Hyperspectral Imaging Based on Dimension-Discriminative Low-Rank Tensor Recovery

lambda-Net: Reconstruct Hyperspectral Images From a Snapshot Measurement

Detecting and Simulating Artifacts in GAN Fake Images

Hyperspectral Image Reconstruction Using a Deep Spatial-Spectral Prior

Explicitizing an Implicit Bias of the Frequency Principle in Two-layer Neural Networks

HyperReconNet: Joint Coded Aperture Optimization and Image Reconstruction for Compressive Hyperspectral Imaging

Spectral-Spatial Attention Networks for Hyperspectral Image Classification

Objects as Points

Structure Tensor and Guided Filtering-Based Algorithm for Hyperspectral Anomaly Detection

Frequency Principle: Fourier Analysis Sheds Light on Deep Neural Networks

Efficient Attention: Attention with Linear Complexities

Gather-Excite: Exploiting Feature Context in Convolutional Neural Networks

Rank Minimization for Snapshot Compressive Imaging

CBAM: Convolutional Block Attention Module

Training behavior of deep neural network in frequency domain

On the Spectral Bias of Neural Networks

Self-Attention with Relative Position Representations

MobileNetV2: Inverted Residuals and Linear Bottlenecks

Non-local Neural Networks

High-quality hyperspectral reconstruction using a spectral prior

Adaptive Nonlocal Sparse Representation for Dual-Camera Compressive Hyperspectral Imaging

Squeeze-and-Excitation Networks

Attention is All you Need

MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications

Xception: Deep Learning with Depthwise Separable Convolutions

Deep Residual Learning for Image Recognition

Generalized alternating projection based total variation minimization for compressive sensing

Adam: A Method for Stochastic Optimization

Coded aperture compressive temporal imaging

3D imaging spectroscopy for measuring hyperspectral patterns on solid objects

Multiframe image estimation for coded aperture snapshot spectral imagers.

Video rate spectral imaging using a coded aperture snapshot spectral imager.

Single disperser design for coded aperture snapshot spectral imaging.

Multispectral Imaging Using Multiplexed Illumination

A New TwIST: Two-Step Iterative Shrinkage/Thresholding Algorithms for Image Restoration

Gradient Projection for Sparse Reconstruction: Application to Compressed Sensing and Other Inverse Problems

Image quality assessment: from error visibility to structural similarity

Face recognition in hyperspectral images

Supplementary Document : Spatial-spectral Encoded Compressive Hyperspectral Imaging

Hyperspectral Face Recognition using 3D-DCT and Partial Least Squares

Representing Scenes as Neural Radiance Fields for View Synthesis

Field of Study

Journal Information

Name

Page

Venue Information

Name

Type

URL

Alternate Names