CBIR Web Interface

A web-based interface for Content-Based Image Retrieval. Upload a query image and find similar images using various feature extraction methods.

Quick Start

Prerequisites

• Python 3.8+ — python --version
• CMake — Download or brew install cmake (macOS) / sudo apt install cmake (Linux)
• OpenCV — Download (Windows) or brew install opencv (macOS) / sudo apt install libopencv-dev (Linux)
• C++ Compiler — Visual Studio (Windows), Xcode (macOS), or GCC (Linux)

You can watch the demo of this project at the link below:

Demo Video:
https://drive.google.com/file/d/1NYAok6uiD5psRPR5hImfyYbnEgWzRcQx/view?usp=sharing

Run

Windows

Double-click setup_and_run.bat or run via command prompt:

git clone https://github.com/dkumar247/cbir.git
cd cbir
setup_and_run.bat

macOS / Linux

git clone https://github.com/dkumar247/cbir.git
cd cbir
chmod +x setup_and_run.sh
./setup_and_run.sh

The script will:

1. Build the C++ backend (using CMake)
2. Install Python dependencies (Flask)
3. Start the web server at http://localhost:5001

---

How to Use

1. Click "Choose Image" and select any .jpg image
2. Select a feature extraction method from the dropdown
3. Set the number of results (1–20)
4. Click "Search Similar Images"

Feature Methods

Method	Description
Baseline (7×7)	Center patch comparison
RGB Histogram	Global color distribution
Multi-Histogram	Top/bottom spatial color matching
Texture + Color	Sobel gradients + RGB histogram
DNN Embedding	ResNet18 semantic similarity

---

Configuration

All paths and settings are in config.py. By default, relative paths are used so everything works out of the box.

---

Project Structure

cbir/
├── setup_and_run.sh        # ← Run this to start
├── config.py               # All configurable settings
├── app.py                  # Flask backend
├── templates/
│   └── index.html          # Web UI
├── cpp_src/                # C++ CBIR engine (built automatically)
├── data/                   # Image database + pre-computed features
└── requirements.txt        # Python dependencies

---

Troubleshooting

Problem	Solution
Problem	Solution
---------	----------
cmake: command not found	Install CMake and ensure it's in your system PATH.
OpenCV not found (Windows)	Set OpenCV_DIR environment variable to your opencv/build folder or run CMake with -DOpenCV_DIR="C:/path/to/opencv/build".
dll missing (Windows)	Add OpenCV bin (containing DLLs) to PATH or copy DLLs next to the executable.
flask not found	Run pip install -r requirements.txt.
Port 5001 in use	(macOS/Linux) lsof -ti:5001 | xargs kill (Windows) Use Task Manager to kill Python processes.

---

Authors

Deepanshu Kumar and Prathmesh Barapatre