Texture Image Classification Using DWT AlexNet Feature Fusion and Deep Neural Networks
A paper proposes a hybrid feature fusion framework, DWT_AlexNet_DNN, combining Discrete Wavelet Transform features with AlexNet deep features for texture image classification.
The paper introduces DWT_AlexNet_DNN, a hybrid framework that fuses Discrete Wavelet Transform (DWT) features with deep features from AlexNet to improve texture image classification. It addresses limitations of handcrafted features in representing complex patterns and deep learning models in exploiting multiscale spatial-frequency information.
The approach integrates DWT-based multiscale spatial-frequency features with AlexNet's learned representations, suggesting a hybrid feature fusion strategy that may enhance classification accuracy for texture images.
Texture classification is relevant to industrial inspection, medical imaging, and remote sensing; hybrid feature fusion could improve robustness in these applications.
Improved texture classification could benefit automated quality control, medical diagnostics, and satellite image analysis, potentially reducing manual inspection costs.
Potential next signals include experimental results on benchmark texture datasets, comparisons with state-of-the-art methods, and possible extensions to other CNN architectures or wavelet transforms.