Hrishikesh Telang

AWS ETL Pipeline Manager for YouTube User Data Analysis

Scalable Data Pipeline for YouTube Insights and Advertising Success

Calorie Advisor: AI-Powered Food Analysis App

Problem Statement

Maintaining a healthy diet is crucial for overall well-being, but it can be challenging to track the nutritional content and caloric intake of daily meals. Many individuals struggle to assess whether their food choices are balanced and healthy, leading to potential health issues over time.

Project Description

Calorie Advisor is a Large Image Model (LIM) application powered by Google Gemini Pro Vision API. This app enables users to analyze their meals by simply uploading an image of their food. The app provides:

A breakdown of the food items and their calorie content.
An assessment of whether the meal is healthy or not.
Insights into the macronutrient distribution (carbohydrates, fats, fiber, sugar, etc.).

The goal of this app is to encourage healthier food choices by providing actionable nutritional insights in real time.

Methodology and Steps

1. Problem Understanding and Objective Setting

Identify the challenge of tracking nutrition and caloric intake from meals.
Define the app’s objectives to provide detailed and accurate food analysis.

2. Application Architecture

Frontend: Built using Streamlit for an intuitive interface where users can upload food images.
Backend: Leveraged Google Gemini Pro Vision API for image analysis and nutritional insights.

3. Steps to Build

Environment Setup: Created a virtual environment and configured dependencies.
Input Handling: Enabled users to upload images and converted the input into a format compatible with the Google Gemini Pro Vision API.
Custom Prompt Design: Developed a custom prompt to instruct the model to analyze images for food items, calorie counts, and nutritional balance.
Data Processing: Integrated the Vision API to retrieve detailed insights based on the uploaded food images.
Result Display: Presented calorie breakdown and health assessment on the app’s interface.

Tools and Technologies

Models: Google Gemini Pro Vision API
Frameworks: Streamlit
Libraries: google-generative-ai, python-dotenv, Pillow (for image handling)
Development Environment: Python (3.10), Conda for virtual environment management

Challenges Faced

Prompt Engineering: Crafting an input prompt to ensure the Google Gemini Pro Vision API provided detailed and relevant responses.
Data Format Conversion: Adapting the uploaded image to the format required by the Vision API.
Accuracy: Balancing accuracy in calorie and nutrient breakdowns based on image analysis, which can be limited by the quality and composition of the uploaded image.

Outcome

Developed a fully functional Calorie Advisor app capable of analyzing meal images in real time.
Successfully identified food items, calculated calorie counts, and provided macronutrient breakdowns for uploaded images.
Encouraged healthier eating habits by offering insights into the nutritional quality of meals.

What I Learned

Gained expertise in integrating Google Gemini Pro Vision API for image-based AI applications.
Enhanced skills in prompt engineering for fine-tuning large model responses.
Improved knowledge of building end-to-end applications using Streamlit.
Strengthened understanding of food nutrition and calorie tracking from a technical perspective.

Future Scope

Multi-Language Support: Enable analysis for meals labeled in different languages.
Expanded Database: Incorporate a larger database for recognizing more food items and cuisines.
Mobile Integration: Develop a mobile version of the app for easier access and usage.

How to Run the Application

Set up the Environment: ```bash conda create -n calorie-advisor python=3.10 conda activate calorie-advisor pip install -r requirements.txt
Create the .env File: ```bash Add your Google API key: GOOGLE_API_KEY=your_google_api_key_here
Run the App: ```bash streamlit run app.py
Upload a Food Image: ```bash Upload an image of your meal to analyze its nutritional content.