Hybrid CNN Bi-LSTM neural network for Hyperspectral image classification
Alok Ranjan Sahoo, Pavan Chakraborty
TL;DR
The paper targets hyperspectral image classification by addressing high dimensionality and inter-layer information loss. It introduces HSSNB, a hybrid network that stacks 3D CNN, 2D CNN, and Bi-LSTM, with PCA-based spectral reduction and 25×25 windows to efficiently learn spectral–spatial features and inter-layer correlations. Compared to state-of-the-art models, HSSNB achieves competitive or superior accuracy (e.g., OA of 100% on SA and ~99.8% on IP/PU) using about 70% fewer trainable parameters, and maintains strong performance with as little as 10% training data. This approach offers a parameter-efficient alternative for hyperspectral classification with practical benefits for remote sensing applications.
Abstract
Hyper spectral images have drawn the attention of the researchers for its complexity to classify. It has nonlinear relation between the materials and the spectral information provided by the HSI image. Deep learning methods have shown superiority in learning this nonlinearity in comparison to traditional machine learning methods. Use of 3-D CNN along with 2-D CNN have shown great success for learning spatial and spectral features. However, it uses comparatively large number of parameters. Moreover, it is not effective to learn inter layer information. Hence, this paper proposes a neural network combining 3-D CNN, 2-D CNN and Bi-LSTM. The performance of this model has been tested on Indian Pines(IP) University of Pavia(PU) and Salinas Scene(SA) data sets. The results are compared with the state of-the-art deep learning-based models. This model performed better in all three datasets. It could achieve 99.83, 99.98 and 100 percent accuracy using only 30 percent trainable parameters of the state-of-art model in IP, PU and SA datasets respectively.