This project involved the development of an AI-driven application designed to detect and identify various plant diseases through image recognition. The core objective was to build a functional system that can analyze photos of plants and provide an immediate diagnosis based on visual patterns. By utilizing a custom-trained neural network, the platform offers a streamlined way to recognize specific botanical pathologies with high accuracy.
I developed the machine learning model and the frontend of the application. My work included preprocessing the image datasets for training and designing the user interface. I was responsible for integrating the detection logic with the web interface to ensure the transition from image upload to results worked correctly.
The development began by training a custom model in Teachable Machine to categorize different plant diseases based on image datasets. To make this model functional within a professional environment, I exported it and used TensorFlow to handle the backend logic and image processing. The final application was built using Python and Streamlit, which allowed us to deploy the model as an interactive web tool where users can upload photos for instant analysis.
Working on this project provided me with practical experience in training and deploying deep learning models within a web-based ecosystem. I improved my Python proficiency and gained a solid understanding of how to use frameworks like Streamlit to turn data scripts into interactive applications.