NutriInsightX

AI-Powered Food Label Analysis & Nutrition Intelligence System

---

Overview

NutriInsightX is an AI-powered food label analysis platform that combines Optical Character Recognition (OCR), Machine Learning, and nutritional risk assessment to help users understand packaged food products.

The system extracts ingredient information from food labels, detects allergens and additives, evaluates health risks, and predicts Nutri-Score grades using a Random Forest classifier trained on the OpenFoodFacts Kaggle dataset.

---

Key Achievements

190,969 Training Samples

94.54% Model Accuracy

7 Nutritional Features

A-E Nutri-Score Classes

---

Machine Learning Pipeline

flowchart TD

A["📦 OpenFoodFacts Dataset
190,969 Records"]

--> B["🧹 Data Cleaning"]

--> C["⚙️ Feature Selection"]

--> D["📊 Nutritional Features
Energy
Fat
Saturated Fat
Sugar
Protein
Fiber
Salt"]

--> E["✂️ Train-Test Split
80% / 20%"]

--> F["🌲 Random Forest Training"]

--> G["🔍 Hyperparameter Tuning"]

--> H["📈 Model Evaluation"]

--> I["🎯 Accuracy
94.54%"]

--> J["🏆 Nutri-Score Prediction
A • B • C • D • E"]

Loading

---

Dataset

Source

OpenFoodFacts (Kaggle)

Records Used

190,969 Food Products

Target Variable

nutrition_grade_fr

Features

energy_100g
fat_100g
saturated-fat_100g
sugars_100g
proteins_100g
fiber_100g
salt_100g

---

Model Performance

Metric	Value
Dataset	OpenFoodFacts
Records	190,969
Algorithm	Random Forest
Accuracy	94.54%
Target	nutrition_grade_fr
Classes	A, B, C, D, E

Classification Performance

Grade	Precision	Recall	F1-Score
A	0.97	0.95	0.96
B	0.90	0.91	0.91
C	0.93	0.94	0.94
D	0.96	0.96	0.96
E	0.97	0.94	0.96

---

Features

OCR Food Label Analysis Image Upload Ingredient Extraction Food Label Processing Real Package Recognition	Nutri-Score Prediction Random Forest Classifier Grade Prediction A-E Nutritional Feature Analysis ML-Based Classification
Allergen Detection Milk Soy Wheat Gluten Peanuts Tree Nuts	Health Risk Analysis Sugar Assessment Salt Assessment Saturated Fat Analysis Energy Evaluation
Additive Detection Food Additives Preservatives Ingredient Screening Risk Awareness	Personalized Recommendations Consumer Guidance Allergy Warnings Risk Insights Nutritional Suggestions

---

System Architecture

Food Package Image
        │
        ▼
     EasyOCR
        │
        ▼
 Ingredient Extraction
        │
 ┌──────┼────────┐
 ▼      ▼        ▼
Language Allergens Additives
Detect   Detect    Detect
        │
        ▼
 Health Risk Engine
        │
        ▼
 Random Forest Model
        │
        ▼
 Nutri-Score Prediction
        │
        ▼
 Recommendations

---

Technology Stack

Layer	Technologies
User Interface	Streamlit
Computer Vision	EasyOCR, OpenCV, Pillow
Machine Learning	Scikit-Learn, Random Forest
Data Processing	Pandas, NumPy
Language Detection	LangDetect
Visualization	Streamlit Charts, Matplotlib
Dataset	OpenFoodFacts (Kaggle)
Programming Language	Python 3.11

---

Application Preview

Light Theme

---

Dark Theme

---

OCR Extraction

---

Health Risk Analysis

---

Nutri-Score Prediction

---

Project Structure

NutriInsightX
│
├── app.py
├── preprocess.py
├── train_model.py
│
├── src
│   ├── ocr_engine.py
│   ├── allergen_detector.py
│   ├── additive_detector.py
│   ├── health_risk.py
│   ├── language_detector.py
│   ├── recommender.py
│   └── nutriscore_predictor.py
│
├── knowledge_base
│   ├── allergens.json
│   └── food_additives.json
│
├── screenshots
│
├── dataset
│
└── models

---

Installation

git clone https://github.com/LovelySharma-dev/NutriInsightX.git

cd NutriInsightX

pip install -r requirements.txt

streamlit run app.py

---

Future Enhancements

• SHAP Explainability Integration
• Barcode Scanner Support
• Real-Time Camera OCR
• Automated Nutrition Extraction
• Product Comparison Engine
• Mobile Application

---

License

This project is licensed under the MIT License.

See the LICENSE file for details.

---

Final Year Report

AI-Powered Food Label Analysis Using OCR and Machine Learning