This paper presents a novel approach to texture analysis and classification using a tree-structured wavelet transform (TSWT). The TSWT is an extension of the conventional pyramid-structured wavelet transform (PSWT) that allows for more flexible decomposition in frequency channels, particularly in the middle frequency range where natural textures are often found. The TSWT is motivated by the observation that natural textures are quasi-periodic signals with dominant frequencies in the middle frequency channels. The paper introduces a progressive texture classification algorithm that is both computationally efficient and performs well. The performance of the TSWT is compared with other techniques such as DCT, DST, DHT, Gabor filters, and Laws filters. Experimental results show that the TSWT outperforms these methods, especially for textures with significant middle frequency channels. The paper also discusses the sensitivity of the TSWT to noisy data and its performance on visually similar textures, demonstrating its robustness and effectiveness. The authors conclude that the TSWT is a powerful tool for texture analysis and classification, offering advantages over traditional methods in terms of frequency resolution and adaptability.This paper presents a novel approach to texture analysis and classification using a tree-structured wavelet transform (TSWT). The TSWT is an extension of the conventional pyramid-structured wavelet transform (PSWT) that allows for more flexible decomposition in frequency channels, particularly in the middle frequency range where natural textures are often found. The TSWT is motivated by the observation that natural textures are quasi-periodic signals with dominant frequencies in the middle frequency channels. The paper introduces a progressive texture classification algorithm that is both computationally efficient and performs well. The performance of the TSWT is compared with other techniques such as DCT, DST, DHT, Gabor filters, and Laws filters. Experimental results show that the TSWT outperforms these methods, especially for textures with significant middle frequency channels. The paper also discusses the sensitivity of the TSWT to noisy data and its performance on visually similar textures, demonstrating its robustness and effectiveness. The authors conclude that the TSWT is a powerful tool for texture analysis and classification, offering advantages over traditional methods in terms of frequency resolution and adaptability.