SIMPD V1: South Indian Medicinal Plants dataset (Version 1)
Description
Automated plant recognition in real-time, wild settings presents a significant challenge in both botanical taxonomy and computer vision. Existing databases of medicinal plants are typically designed for specific plant classes and are often created in controlled laboratory environments with plain white backgrounds. Traditional handcrafted feature engineering is inadequate for handling large-scale datasets consisting of unconstrained images. Furthermore, existing datasets lack mobile-based plant images captured in natural scenes, and they exhibit significant variability depending on contributors, camera types, geographic locations, seasonal variations, and individual plant characteristics. To address this gap, a new database of medicinal plants found in the wild has been developed. This dataset, named South Indian Medicinal Plants (SIMP) Dataset Version 1, comprises 2,503 images collected from various locations in and around Tirunelveli. The SIMP dataset includes 20 different classes of medicinal plants, encompassing herbs, shrubs, creepers, climbers, wild plants, and trees.