Texture Classification Using Deep Convolutional Neural Networks

Texture classification is an essential task in computer vision and image processing, commonly applied in areas like image recognition and remote sensing. Traditional methods such as Local Binary Pattern (LBP) have limitations in capturing complex texture features. This paper proposes the use of deep convolutional neural networks (CNNs) for texture classification, which can extract complicate texture details and significantly improve classification accuracy. Experimental results on the KTH - TIPS database confirm the superiority of the proposed CNN - based method over conventional LBP technique.