From Frisee to Belgian Endive: A Deep Learning and Ensemble Approach to Classifying Endive Varieties in India

The proposed study looks at the types of Frisée, Escarole, and Belgian Endive in India using Convolutional Neural Networks (CNN) with team learning. A set with 9,000 big pictures was carefully gathered. It showed many different features of these endive types. The study works on making models better and faster in telling the difference between complex endive types. This is important for farming practices and selling them at markets. The made model got a very good total correctness of 98.17% in sorting the different types of endive. The table with details shows that this model can accurately sort every kind of endive. It uses metrics like precision, recall, and F1-score to measure its performance. Figure 3 shows a table called the confusion matrix. It helps people understand why certain predictions are wrong and gives tips on how to make them better for more precise results. This study is a big step forward to make sorting endive types automatic in India, reducing the need to depend on human opinions. The model's high precision promises better farming methods, making sure food quality is safe and simplifying selling in markets. Furthermore, it can do more than just categorize endives. It also has the potential to help sort out plants and make farming machines work better. This study shows how CNN and Ensemble Learning can change farming by using smart computer methods. Using these techniques correctly shows what can be done in the future. They aim to improve accuracy, speed, and long-lasting methods of farming not only in India but also around the world.