🍓 Ichigo: Intelligent Multi-Fruit Classification and Quality Analysis System

📌 Introduction

Traditional fruit classification methods are labor-intensive and error-prone. Ichigo is an AI-driven system that utilizes deep learning models like CNN, ResNet-50, and VGG-16 to classify fruits as fresh or rotten. By integrating machine learning with advanced image processing, Ichigo enhances food quality control, reduces waste, and optimizes supply chains. Designed to be scalable and accessible, it serves small farms and large food industries alike.

❗ Problem Statements

Despite advances in computer vision and AI, challenges persist in fruit classification:

🔸 Manual inspection is prone to inconsistencies – Human judgment can be subjective, leading to quality control issues.
🔸 Limited datasets hinder classification accuracy – AI models struggle with underrepresented fruit types.
🔸 Scalability for real-time processing – Many systems fail to handle large-scale, high-speed operations efficiently.

🎯 Objectives

The Ichigo system aims to:

✅ Enhance classification accuracy using deep learning.
✅ Enable real-time quality assessment for efficient sorting.
✅ Ensure scalability for farms, supermarkets, and industries.

🔥 System Features

🚀 Deep Learning-Based Classification – Uses CNN, ResNet-50, and VGG-16.
🖥️ Web-Based Interface – Built with Flask, HTML, CSS, and JavaScript.
📊 Data Preprocessing Techniques – Image enhancement, sharpening, and edge detection.
🔍 Real-Time Object Detection – Segmentation and bounding box.

📊 Model Performance

📌 Public Dataset Performance (Without Preprocessing)

Model	Accuracy (%)	Precision (%)	Recall (%)	F1-Score (%)
CNN	72.50	74.50	72.75	70.84
ResNet50	98.75	98.76	98.71	98.72
VGG16	97.03	97.08	96.92	96.96

📌 Public Dataset Performance (With Preprocessing)

Model	Accuracy (%)	Precision (%)	Recall (%)	F1-Score (%)
CNN	72.81	73.01	72.70	71.88
ResNet50	99.22	99.21	99.20	99.20
VGG16	97.81	97.69	97.68	97.65

🔗 Public Dataset Used

Sultana, Nusrat; Jahan, Musfika; Uddin, Mohammad Shorif (2022), “Fresh and Rotten Fruits Dataset for Machine-Based Evaluation of Fruit Quality”, Mendeley Data, V1, doi: 10.17632/bdd69gyhv8.1

📌 Self-Collected Dataset Performance (Without Preprocessing)

Model	Accuracy 1 (%)	Accuracy 2 (%)	Average (%)
CNN	26.00	24.27	25.14
ResNet50	71.00	78.64	74.82
VGG16	52.50	55.83	54.17

📌 Self-Collected Dataset Performance (With Preprocessing)

Model	Accuracy 1 (%)	Accuracy 2 (%)	Average (%)
CNN	18.00	12.62	15.31
ResNet50	53.00	55.83	54.42
VGG16	47.50	58.25	52.88

🔗 Self-Collected Dataset Used (Dataset 1)

Refer to the end of this README.

🛠️ Installation Guide

📌 Prerequisites

Ensure you have the following installed:

• 🐍 Python 3.9+

⚙️ Steps to Install and Run

1️⃣ Clone the repository 🖥️

git clone https://github.com/rydzze/Ichigo-Fruit-Classifier.git
cd Ichigo  

2️⃣ Install dependencies 📦

pip install -r requirements.txt  

3️⃣ Run the application

python run.py  

4️⃣ Access the system: 🌍

http://localhost:686  

📸 Screenshots of User Interface

🏆 Contribution

We would like to express our gratitude to the following individuals for their contributions to Ichigo:

• Muhammad Ariff Ridzlan
• Muhammad Hafiz
• Siti Nur Aisyah
• Nurul Hurul Aini

Your dedication and expertise have been instrumental in the development of this system. 🚀💡

💻 Google Drive Link (Alternative)

Ichigo, including the self-collected dataset and .h5 model files.