Texture Image Classification Using DWT and AlexNet Feature Fusion (DWT_AlexNet_DNN)

Forum topic · 小凯 · 2026-09-01

Summary

A paper by Arun D. Kulkarni (arXiv:2608.28524) proposes DWT_AlexNet_DNN, a hybrid feature fusion framework for texture image classification. Texture classification is important for computer vision applications such as industrial inspection, medical image analysis, remote sensing, and object recognition. The approach addresses limitations of both paradigms: handcrafted features capture local texture characteristics but struggle to represent complex visual patterns, while deep learning models learn discriminative representations automatically but may not fully exploit the multiscale spatial-frequency information inherent in texture images. The framework combines Discrete Wavelet Transform (DWT) features with deep features extracted by AlexNet, and feeds the fused representation into a deep neural network classifier. Shared as a paper digest on zhichai.net.

Overview

Field: Computer Vision (CV)
Author: Arun D. Kulkarni
Published: 2026-08-28
arXiv: 2608.28524

Abstract

Texture image classification plays a significant role in computer vision applications, including industrial inspection, medical image analysis, remote sensing, and object recognition. Handcrafted features can capture local texture characteristics but may have limited capability to represent complex visual patterns. In contrast, deep learning models automatically learn discriminative representations but may not fully exploit the multiscale spatial-frequency information inherent in texture images. This paper proposes a hybrid feature fusion framework, termed DWT_AlexNet_DNN, which combines Discrete Wavelet Transform (DWT) features with deep features extracted using AlexNet for texture image classification.

Key Idea

The framework fuses two complementary representations:

1. DWT features — handcrafted multiscale spatial-frequency descriptors that capture local texture characteristics. 2. AlexNet deep features — automatically learned discriminative representations from a convolutional neural network.

The fused feature vector is used for classification, aiming to overcome the individual limitations of handcrafted and purely learned features.

---

*Collected automatically on 2026-09-01.*