Vastradarpana A Novel Method to Detect Saree Using Image Processing Techniques

Vastradarpana A Novel Method to Detect Saree Using Image Processing Techniques

Mr. Yogesh B S, Nisarga K, Nithya M, Saijyoti G Meti, Sanjana H Goud

Student, Information Science & Engineering

Dayananda Sagar college of Engineering, Bengaluru

Abstract-Identifying authenticated sarees through human vision and feel is a challenging aspect. Secondly Saree industry is multibillion businesses in Indian Industry. Each silk saree cost range from Rs 10000 to Rs 500000. Every consumer tries to identify authenticated saree before purchasing. With today’s technology changes we have prepared novel approach to detect the authentication of the saree. This work offers a method for classifying and detecting saree types using computer vision techniques. The visual features of sarees are classified by machine learning and image processing; preprocessing improves the classification model. With the use of region based CNNs (RCNN) extract texture features that are essential for differentiating Kanchivaram Silk Sarees. The model ensures stability across datasets and training conditions, as demonstrated by experimental findings that show its improved accuracy and speed over traditional deep learning techniques. This research increases computer vision in apparel recognition and categorization with applications in e-commerce and fashion recommendation.

Keywords-Sarees, Kanchivaram, Rcnn, Tensorflow