🖼️ Images to Insights

Overview

Images to Insights is a powerful visual intelligence tool that detects and summarizes changes between two images using a hybrid approach combining:

• YOLOv8 (object-level difference detection)
• OpenCV (pixel-level difference detection)

Features

• ✅ YOLOv8-Based Object Comparison

  • Detects object-level differences (e.g., cars, people, etc.)
  • Identifies added, removed, or moved entities

• 🔍 Pixel-Level Comparison with OpenCV

  • Computes and visualizes raw pixel differences
  • Highlights subtle changes missed by object detectors

• 🧾 Natural Language Summary

  • Generates human-readable summaries of what changed
  • Supports both object and pixel difference explanations

• 💻 Streamlit UI

  • Simple web interface to upload and compare images
  • View both bounding boxes and pixel diffs side-by-side

Packages used for this project:

• ultralytics for YOLOv8 model(object level detection)
• opencv-python(for image processing and pixel level detection
• numpy for converting images to arrays
• streamlit for interface
• Pillow for image handling, format conversion

Example Outputs

Below are sample visual outputs generated by the system:

• YOLO-based Entity Detection

  

• OpenCV-based Pixel Difference

  

---

User Interface

The following screenshot shows the web interface built using Streamlit:

Installation

1. Clone the repository:

   git clone https://github.com/PhoenixAlpha23/image-to-insights.git
   cd image-to-insights

2. Install required dependencies:

   pip install -r requirements.txt

3. *Set Up Groq API Key Before running the app, export your Groq API key as an environment variable:

   export GROQ_API_KEY=your_api_key_here
   

4. Run the application:

   streamlit run app.py