Hyperspectral Image Classification Using Deep Pixel-Pair Features

By Wei Li, Guodong Wu, Fan Zhang, Qian Du. IEEE Transactions on Geoscience and Remote Sensing, 2017

714 citations recorded by the source corpus.

Subfields: Atmospheric Science, Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/d639fc711f5c22da1995c1bc0f24300aa70bf4ec

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Atmospheric Science

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Deep Pyramidal Residual Networks for Spectral–Spatial Hyperspectral Image Classification (2019, methodology): “[44] combined the CNN model with pixel pairs to learn discriminative features, using a majority voting strategy to obtain the final classification result.”
Multiscale Dynamic Graph Convolutional Network for Hyperspectral Image Classification (2019, methodology): “In the recent work of Li et al. [50], deep CNN is used to learn pixel-pair features, and the classiﬁcation results of pixels in different pairs from the neighborhood are then fused.”
Deep learning in remote sensing: a review (2017, methodology): “In [30], to allow a CNN to be appropriately trained using limited labeled data, the authors present a novel pixel-pair CNN to significantly augment the number of training samples.”
Multisource Remote Sensing Data Classification Based on Convolutional Neural Network (2018, methodology): “Note that in Table X, CNN-PPF using the HSI alone achieves an accuracy of 93.28%, which is lower than the one in the literature [39].”
Hyperspectral Image Classification—Traditional to Deep Models: A Survey for Future Prospects (2021, background): “For instance, [149] presented an improved pixel pair feature (PPF) [150] approach called spatial pixel pair feature which is different from traditional PPFs with respect to two main aspects: one is the selection of pixel pair that is only the pixel from the immediate neighborhood of central pixel ca”
Advances in Hyperspectral Image and Signal Processing: A Comprehensive Overview of the State of the Art (2017, background): “, Apache Spark, Caffe, Theano, Torch, and TensorFlow), which also find application in the hyperspectral analysis community [64], [65], [84], [337], [338].”
Capsule Networks for Hyperspectral Image Classification (2019, methodology): “[25] propose the use of pixel-pair features under a CNN-based classification scheme in order to increase the number of training samples and, hence, the resulting classification performance.”
Generative Adversarial Networks for Hyperspectral Image Classification (2018, methodology): “For the PCA-CNN, the CNN is conducted on the three principal components, which is useful when the training samples are limited [20].”
A review of deep learning used in the hyperspectral image analysis for agriculture (2021, background): “Some researchers regard the SAEs as a part of the framework to extract effective features (Chen and Li 2017).”
Spatial-Spectral Transformer for Hyperspectral Image Classification (2021, methodology): “Moreover, an interesting work was proposed in [30], which used CNN to extract pixel-pair features for HSI classification and obtained good classification performance.”
Deep Learning for Hyperspectral Image Classification: An Overview (2019, methodology): “In this paper, we have investigated several recent state-of-the-art deep learning-based approaches, including 3-D-CNN [56], Gabor-CNN [79], CNN with PPFs (CNNPPFs) [61], Siamese CNN (S-CNN) [89], 3-D-GAN [46],”
Deep Few-Shot Learning for Hyperspectral Image Classification (2019, methodology): “The compared methods include CNN [37], R-PCA-CNN [34], CNN-PPF [45], and C-CNN [47].”
Diverse Region-Based CNN for Hyperspectral Image Classification (2018, methodology): “CNN-PPF [38], CD-CNN [39], and SS-CNN [36].”
Hyperspectral Image Classification With Deep Feature Fusion Network (2018, methodology): “With the strong ability of extracting features, the CNN has been recently introduced to classify HSIs [43], [46], [48].”
Deep Learning Meets Hyperspectral Image Analysis: A Multidisciplinary Review (2019, methodology): “Among them, random pixel-pair features (PPF) was introduced in [21], which exploits the similarity of the pixels of the same class to augment the training data, where a deep CNN with multi layers is then employed to learn these PPF.”
HSI-BERT: Hyperspectral Image Classification Using the Bidirectional Encoder Representation From Transformers (2020, methodology): “[30] proposed a novel CNN-based method to encode the pixel-pair features and make predictions by a voting strategy.”