Detection And Identification of Ethnic Attires Using Deep Learning And GI Tags

Northeast India has a rich cultural heritage, shown through its variety of traditional clothing, each with unique meanings and history. This study aims to make it easier to identify and recognize these traditional outfits using advanced computer techniques, specifically deep learning combined with Geographic Indication (GI) tags. GI tags help connect each attire to its specific cultural and regional background. We used ResNet50 model to build a system that can automatically recognize and categorize traditional clothing from Northeast India. By adding GI tags, our system can more accurately identify these outfits based on their cultural origins. MobileNetV3 and EfficientNetB0 models were used to compare with the ResNet50 accuracy. We have tested this system with collection of ethnic dress images and found that it effectively distinguishes between different dress types with high accuracy. Using deep learning with GI tags not only improves the accuracy of identifying traditional outfits but also helps preserve and promote the cultural heritage of Northeast India. This work highlights how advanced technology can support the documentation and appreciation of unique cultural traditions, providing a useful tool for cultural preservation.