Content-Aware Image Resizer

An advanced content-aware, image resizing program utilizing seam carving algorithm in C++ that allows image to be scaled without losing/distorting meaningful image landmarks. This is implemented according to the seam carving algorithm described in Avidan & Shamir, 2007 which finds the path of least energy density for removal by computing a energy matrix mapping to the image.

Features

• Seam Carving Algorithm: Intelligently resizes images by removing or adding seams based on content importance.
• Customizable Energy Functions: Utilizes Sobel filtering for edge detection; easily extendable for custom energy functions.
• Command-Line Interface: Provides flexibility for batch processing and integration into workflows.
• Visualization of Seam Removal: Optionally displays the seam removal process for better understanding.

Installation

Prerequisites

• C++ compiler supporting C++11 standard
• openCV

Building the Project

1. Clone the repository:

   git clone https://github.com/felixlu4725/Content-Aware-Image-Resizer.git
   cd Content-Aware-Image-Resizer
   

2. Compile the code:

   g++ -std=c++11 -o seam_carving main.cpp -lopencv_core -lopencv_highgui -lopencv_imgcodecs -lopencv_imgproc
   

   Ensure that OpenCV is correctly installed and linked.

Usage

./seam_carving <input_image> <output_image> <new_width> <new_height> [display]

• <input_image>: Path to the input image file.
• <output_image>: Path to save the resized image.
• <new_width>: Desired width of the output image.
• <new_height>: Desired height of the output image.
• [display] (optional): Set to 1 to display the seam removal process; defaults to 0.

Example

./seam_carving input.jpg output.jpg 800 600 1

This command resizes input.jpg to 800x600 pixels and displays the seam removal process.

How It Works

1. Compute Energy Map: Calculates the energy of each pixel using the Sobel operator to detect edges.
2. Find Optimal Seams: Applies dynamic programming to identify the seams with the lowest energy.
3. Remove or Insert Seams: Adjusts the image dimensions by removing or inserting seams based on the desired size.
4. Output Processed Image: Saves and optionally displays the resized image.

By Felix Lu felixlu@umich.edu