A multi-format open-source historic wall image dataset for architects, historians, data scientists to detect, classify, and analyze.
Description
The historical place dataset presents a collection of images of different features gathered from Panam City, a United Nations Educational, Scientific, and Cultural Organization (UNESCO) World Heritage Site situated in Sonargoan, Dhaka, Bangladesh, located at the geographical coordinates 23.6421599°N and 90.6023361°E. The data collection period, from October 6 to October 7, 2023, resulted in a dataset encompassing a total of 2,292 images. The dataset is categorized into five classes: Artistic, Corroded Brick, Corroded Plastic, Fungus, and Living Plant. The data collection process was conducted on a sunny day, ensuring natural lighting conditions, with temperatures ranging from approximately between 26-28°C. The images were captured using the high-resolution camera of an iPhone. The original images were captured in JPEG format with a resolution of 3024 x 4032 pixels and subsequently converted to a resolution of 1080 x 1440. The iPhone 13 was employed for capturing the dataset's photographs. Furthermore, the images have been annotated based on their features, facilitating semantic segmentation, which has garnered considerable attention among data scientists.