Hybrid Fully Connected Tensorized Compression Network for Hyperspectral Image Classification

Li, Heng-Chao; Lin, Zhongqiu; Ma, Tian-Yu; Zhao, Xi-Le; Emery, William J.

doi:10.1109/tgrs.2023.3241193

Cited by 7 publications

(4 citation statements)

References 66 publications

(74 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To increase spectral image classification accuracy and efficacy, researchers have been looking into various parameter performance improvement strategies. One popular strategy is model compression [30], [31], which allows for a reduction in the number of network parameters by using methods such as pruning [32], [33], quantization [34], and decomposition [35]. However, this method results in information loss, which lowers the performance of the model in classification tasks.…”

Section: B Parameter Performance Optimization Methods In Hsi Classifi...mentioning

confidence: 99%

LSSMA: Lightweight Spectral–Spatial Neural Architecture With Multiattention Feature Extraction for Hyperspectral Image Classification

Ding,

Ruan,

Yang

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Deep learning has been utilized for hyperspectral image (HSI) classification in recent years, with notable performance improvements. In particular, convolutional neural networks (CNNs) methods have achieved major advancements in this area. However, there are some drawbacks to the existing CNN-based HSI classification approaches: 1) the lack of effective and simple feature representations in CNNs, which overlook the effects of spectral differences and spatial contextual information; 2) the classification model has an enormous network complexity as a result of its numerous training parameters and high computational requirements; and 3) the category samples in HSI data exhibit a significant long tail distribution issue, which affects the classification performance of HSI. To address these issues, we propose a lightweight spectral-spatial neural architecture with multi-attention feature extraction (LSSMA) for HSI classification. The main work consists of three areas:1) a spectral feature extraction and fusing module (SFEF) is created to facilitate the fusing of spectral-spatial features while reducing the number of trainable parameters and computational complexity of the model. This module uses convolutions of various kernel sizes for residual connection and feature fusion, and introduces group convolution to achieve efficient feature representation; 2)to utilize the spectralspatial correlation of HSI data to its fullest, a multi-scale convolutional activation guided attention mechanism (MsCA) is designed and the position attention module (PAM) is referenced, which can capture spectral differences and spatial contextual relationships between ground objects; and 3) focal loss (FL) is applied in computer vision tasks to the LSSMA model to enhance its capacity to handle category imbalance. Experimental results on four publicly available hyperspectral datasets show that the method obtains better classification performance at a lower computational cost.

show abstract

Section: B Parameter Performance Optimization Methods In Hsi Classifi...mentioning

confidence: 99%

LSSMA: Lightweight Spectral–Spatial Neural Architecture With Multiattention Feature Extraction for Hyperspectral Image Classification

Ding,

Ruan,

Yang

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Simplification methods of the model are mainly divided into model compression and lightweight model design. Li et al [37] proposed a compression network considering the high dimensionality of HSI. A fast and compact 3-D-CNN with few parameters was developed in [38].…”

Section: Hsi Classification Methodsmentioning

confidence: 99%

Network Collaborative Pruning Method for Hyperspectral Image Classification Based on Evolutionary Multi-Task Optimization

et al. 2023

View full text Add to dashboard Cite

Neural network models for hyperspectral images classification are complex and therefore difficult to deploy directly onto mobile platforms. Neural network model compression methods can effectively optimize the storage space and inference time of the model while maintaining the accuracy. Although automated pruning methods can avoid designing pruning rules, they face the problem of search efficiency when optimizing complex networks. In this paper, a network collaborative pruning method is proposed for hyperspectral image classification based on evolutionary multi-task optimization. The proposed method allows classification networks to perform the model pruning task on multiple hyperspectral images simultaneously. Knowledge (the important local sparse structure of the network) is automatically searched and updated by using knowledge transfer between different tasks. The self-adaptive knowledge transfer strategy based on historical information and dormancy mechanism is designed to avoid possible negative transfer and unnecessary consumption of computing resources. The pruned networks can achieve high classification accuracy on hyperspectral data with limited labeled samples. Experiments on multiple hyperspectral images show that the proposed method can effectively realize the compression of the network model and the classification of hyperspectral images.

show abstract

“…Guo et al 52 proposed a contrastive learning method for the compression task. In another work, Li et al 53 introduced a fully connected approach considering the hyperspectral datacube in terms of tensor.…”

Section: Tensor Ring Decompositionmentioning

confidence: 99%

Unmixing aware compression of hyperspectral image by rank aware orthogonal parallel factorization decomposition

Das,

Ghosal

2023

J. Appl. Rem. Sens.

View full text Add to dashboard Cite

.Efficient compression is pertinent for the convenient storage, transmission, and processing of modern high-resolution hyperspectral images (HSI). We propose a high-performance HSI compression method using library-based spectral unmixing and tensor decomposition. Unlike the existing approaches, our proposed work incorporates unmixing in the compression framework and achieves significantly higher compression performance with negligible loss. The proposed library-based unmixing method includes an index for accurate endmember number estimation, followed by exact library pruning and a sparsity regularized formulation with norm-smoothing to compute the abundance maps. As the spectral library is available at the reconstruction (decoder) side also; compressing the abundance maps is as good as compressing the original HSI data. Since the abundance constraints used for the unmixing indicate the correlation of the abundance maps, compressing all abundance maps seems to cause redundant computation. A metric using the image smoothness and information measures is used here to identify the abundance map hardest to compress and the remaining part is left uncompressed. Subsequently, the work compresses the abundance map tensor using orthogonal PARAFAC decomposition with optimal rank determination. The orthogonalization process ensures that the factors span independent subspaces and reduces redundancy, whereas the rank selection prevents noisy or insignificant components. Extensive experiments are carried out to demonstrate that the unmixing workflow leads to negligible loss due to accurate endmember number estimation, exact library pruning, and accurate physically meaningful sparse inversion. Comparative assessments of compression efficacy suggest that the proposed work corresponds to better compression performance and higher classification accuracy.

show abstract

Hybrid Fully Connected Tensorized Compression Network for Hyperspectral Image Classification

Cited by 7 publications

References 66 publications

LSSMA: Lightweight Spectral–Spatial Neural Architecture With Multiattention Feature Extraction for Hyperspectral Image Classification

LSSMA: Lightweight Spectral–Spatial Neural Architecture With Multiattention Feature Extraction for Hyperspectral Image Classification

Network Collaborative Pruning Method for Hyperspectral Image Classification Based on Evolutionary Multi-Task Optimization

Unmixing aware compression of hyperspectral image by rank aware orthogonal parallel factorization decomposition

Contact Info

Product

Resources

About