Access to Ayurvedic texts: This was important in order to find the most relevant information for Ayurvedic ingredients and patients. The database included both classical texts and modern research papers. Availability of a large and diverse dataset of labeled images of medicinal plants and raw materials This was essential for training an accurate machine learning model.
Collecting the appropriate database was a challenge, which we handled by creating our own database of images for medicinal plant identification.

Second was the model training, where we implemented various algorithms and finally succeeded in identifying the medicinal plants with the images accurately. The algorithm is able to identify plants and raw materials accurately, even when faced with noisy or incomplete images. It should also be able to distinguish between similar-looking plants and raw materials. Now the software has a deep understanding of Ayurvedic medicine in order to make accurate recommendations. This includes knowledge of single herbs, minerals, and formulations, as well as the different factors that need to be considered when choosing a remedy.