SAREE TEXTURE ANALYSIS AND CLSSIFICATION USING DEEP LEARNING FRAMEWORK

Authors

  • Saraswathi Author
  • R. Kalavathi Author

Keywords:

Deep learning,, Texture classification, Patch generation, Mask RCNN, VGG-16 network

Abstract

Indian women have worn sarees for centuries and have a significant market share in their clothing. Despite the ability to 
identify saree materials with tactile texture and the possibility to purchase online, it may not always be practicable to touch and feel a 
saree being worn. Moreover, the wide variety of saree materials available in comparison to other types of clothing, as well as the lack 
of experience in recognizing them, make this challenge even more challenging. The rapid advancement of smartphone technology 
makes it possible to capture images of cloth to determine the exact material of a saree and make an online purchase. To overcome the 
aforementioned challenges, a novel deep learning-based saree texture classification framework has been proposed for the rapid 
classification of saree tactile textures. Deep learning-based segmentation is used to detect visual texture and identify sarees from 
collected images. For better understanding the introduction of Mask Region-based Convolutional Neural Network (Mask RCNN) was 
used for generating the saree patches. To categorize saree textures, we propose a deep learning-based VGG-16 network. According to 
the findings, this pipeline performs better than the current deep learning methods and achieves 97.41% accuracy. 

Downloads

Published

2026-06-29