KoyilNet:A Novel Deep Learning Approach for Classifying the Architectural Styles of Tamil Nadu's Heritage Temples
Year: 2026
Type: other
Abstract: The Koyil Heritage Temple Dataset is an image-based dataset developed for the classification of the architectural styles of historical temples in Tamil Nadu. The dataset comprises temple images categorized into five major historical architectural styles/dynasties: Chola, Nayak, Pallava, Pandya, and Vijayanagara. Each image is associated with metadata describing its architectural-style label, deity, geographical location, historical period, and related place/deity information. The dataset contains 892 images in total, distributed across the five architectural-style classes. The class-wise distribution reported in the metadata is: Chola – 246 images, Nayak – 190 images, Pallava – 217 images, Pandya – 237 images, and Vijayanagara – 246 images. Metadata Attributes The dataset metadata includes the following fields: ID – unique identifier for the image/sample. Label_Name – architectural/dynastic class name. Label_Int – numerical class label used for machine-learning classification. Deity – deity/deities associated with the temple. Place – geographical location associated with the temple. Year – historical period or century recorded for the temple. Updated_Place – standardized/updated location information. Common_Deity – standardized deity information. Image Path – storage path corresponding to the temple image
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7204690571