Abstract:
The Identification of Medicinal Plant is important in various fields such as herbal medicine and Ayurvedic field. Accurate identification is required for safe and effective use of medicinal plants in traditional as well as in modern healthcare practices.
However, Manual identification of medicinal plant is very time consuming, requires labours and have errors. In recent years in the advancements of Machine Learning and Computer Vision make it possible to automate the process of identification of different species of medicinal leaves by analysing the input leaf images.
This project presents the comprehensive overview of the Identification of Different Medicinal Plants using Machine Learning techniques and Deep Learning Techniques. The proposed method uses the Indian medicinal leaves dataset from the Kaggle which is the platform for dataset and Machine learning competitions. The dataset contains 80 species of medicinal leaves with approximately 6900 images of leaves in total. The dataset is first pre-processed for image size and scaling then the model was trained using the Tensorflow’s Xception architecture which gave the accurate prediction both in the testing images and new leaves images.
The system’s user-friendly interface allows users to upload the plant leaf image and the system identifies the medicinal plant and shows the result which includes the predicted plant name, botanical name, common names and its medicinal uses. The system has a direct method of searching the medicinal plant uses by giving name of the medicinal plant.
Various stakeholders, including Ayurvedic practitioners who rely on traditional plant therapies and herbal medicine users seeking natural remedies as well as researches and scientists studying the medicinal properties of plant, could be significantly impacted by implementing this project.
Introduction:
Identification of different medicinal plants involves the creation of a system that correctly and accurately identifies the different types of medicinal plants based on their leaves. Traditionally this identification task was done manually which is slow and can have errors. With deep learning, we teach the computers to do it automatically. The problem is to teach the computer program to look at the leaf images and classify them as which species they belong to. This means it must be trained on a large number of image datasets and labeled with their plant names. We also need to ensure that it handles the different backgrounds while identifying. The goal is to make it faster and easier for people to identify the medicinal plant which helps researchers, conservators and people who make herbal medicines
Dataset:
• Dataset Name: Indian Medicinal Leaves
• Dataset Source: The Indian Medicinal Leaves dataset is sourced from Kaggle, a popular platform for datasets and machine learning competitions.
• Dataset Description: It comprises images of medicinal plants, specifically focusing on leaves. The dataset contains a total of 80 species of plants, each represented by a separate subfolder within the dataset directory. There are approximately 6900 images in total across all species, distributed among the respective subfolders based on plant species. The images showcase varying backgrounds, indicating that the dataset captures the plants in diverse environmental settings. This variability in backgrounds enhances the robustness of the dataset, allowing models to generalize better to real-world scenarios.
• Total Images: The dataset contains a total of 6900 images of 80 species.
FrontEnd : HTML ,CSS ,JS
BackEnd: Flask
Install requirement.txt